WO1991018986A1

WO1991018986A1 - Marker for yeast strain identification

Info

Publication number: WO1991018986A1
Application number: PCT/AU1991/000229
Authority: WO
Inventors: Jenny Elizabeth Petering; Paul Anthony Henschke; Peter Langridge
Original assignee: Luminis Pty. Ltd.
Priority date: 1990-05-25
Filing date: 1991-05-27
Publication date: 1991-12-12
Also published as: JPH05506357A; CA2083694A1

Abstract

The present invention provides a novel plasmid comprising a β-glucuronidase (GUS) coding region operably linked to yeast promoter and terminator sequences, whereby said coding region is expressible in yeast. Suitable promoter and terminator sequences are any which allow expression in yeast and are, in particular, Saccharomyces cerevisiae alcohol dehydrogenase promoter and terminator sequences. The plasmid of the present invention is produced by ligating the promoter and terminator sequences to the GUS gene, and inserting the resultant novel construct into an appropriate carrier plasmid. The carrier plasmid is used to introduce the novel construct into a yeast strain by integration into a host chromosome. Assay for GUS activity can then be used to determine the proportion of the marked strain in the total yeast population.

Description

MARKER FOR YEAST STRAIN IDENTIFICATION Abbreviations: GUS S-glucuronidase; MUG 4-methyl umbellif ryl glucuronide; SSPE 0.18M NaCl,0.015M sodium citrate (pH7.0) ,0.001M EDTA; x-gluc 5-bromo-4-chloro-3-indolyl glucuronide.

BACKGROUND TO THE INVENTION Advances in wine fermentations have followed from the application of pure yeast culture inoculation. This practice enables strain differences to be utilized in the production of a wide range of wines, and also provides greater reliability and control of fermentation resulting in wines with fewer flavour defects (1). More recently, the availability of active dried yeasts has given the winemaker even greater scope for exploiting oenological properties of different strains. Raw materials of wine making, however, contain an unknown load of microorganisms, some of which are capable of producing off-flavours and spoiling wine. A suitable method, therefore, to identify and monitor specific wine yeast strains throughout fermentation is necessary for the optimization of inoculation and sterilisation processes.

The potential for genetic markers in yeast strain identification has been recognized and deliberately marked oenological strains were developed by Vezinhet and Lacroix in 1984 (2). An oenological strain, commercialized as 'marked Kl', was developed by selecting for natural mutants in a population of the Lalvin V yeast - it is double marked with diuron and erythromycin resistance genes. While studies involving marked strain Kl have provided an insight into the kinetics of yeast populations during fermentation (3,4), a limitation still exists in that wine yeast strains of choice cannot be easily marked.

A procedure is described here which utilizes recombinant DNA technology to introduce the E. coli^-glucuronidase (GUS) gene as a marker into any desired yeast strains. The GUS gene was developed as a reporter gene system for use in nematodes and, more recently, in the study of plant gene expression (5,6). The advantages of the GUS system as a marker in yeast strains include its low background levels in Saccharomyces cerevisiae and its ease of assay by fluorimetry, spectrophotometry and agar plate tests.

SUMMARY OF THE INVENTION

The present invention provides a novel plasmid comprising a ^-glucuronidase (GUS) coding region operably linked to yeast promoter and terminator sequences, whereby said coding region is expressible in yeast. Suitable promoter and terminator sequences are any which allow expression in yeast and are, in particular, Saccharomyces cerevisiae alcohol dehydrogenase promoter and terminator sequences.

The plasmid of the present invention is produced by ligating the promoter and terminator sequences to the GUS gene, and inserting the resultant novel construct into an appropriate carrier plasmid. Preferably, the carrier plasmid comprises a DNA sequence which is substantially identical to a DNA sequence on the host yeast chromosome, thereby assisting integration of the novel construct into that chromosome. A suitable carrier plasmid is pWX509, comprising the SMR1-410 gene. The SMR1-410 gene has two functions: (a) it serves as a selection marker, and (b) being almost identical in sequence to the yeast ilv 2 gene, it enhances integration of the novel construct into the host chromosome.

The carrier plasmid is used to introduce the novel construct into a yeast strain (e.g. an industrial or polyploid yeast strain, in particular a wine yeast strain) by integration into a host chromosome. Assay for GUS activity can then be used to determine the proportion of the marked strain in the total yeast population.

Most substrates for the GUS enzyme are incapable of crossing the yeast cell membrane, by natural transport. Therefore, the present assay system comprises : (a) adding a substrate for the GUS enzyme (e.g. x-gluσ or MUG) to a sample of a yeast strain; (b) inducing permeation in the cell membrane of the yeast strain; (c) incubating the yeast strain for sufficient time to form detectable quantities of a product of the GUS enzyme reaction; and (d) testing for said product.

The GUS gene is used as a marker gene to monitor yeast strains, particularly in the field of industrial alcoholic fermentation processes.

In one embodiment of the invention, the GUS gene is used to monitor killer yeast strains in mixed culture ferments.

It has been found that a marked yeast killer strain can be used in a mixed culture inoculum to quantify directly the effect of killer toxin on a sensitive yeast strain under fermentation conditions. As a wide range of yeast strains can be readily marked, this system of analysis is unlimited in application and furthermore provides a simple and unequivocal means of quantifying killer yeast strains in mixed culture ferments.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrate the construction of the GUS vector. Figure 2 is a map of pAW220.

Figure 3 illustrates the results of agar plate assay for killer activity. The agar (ph 4.2, 0.003% methylene blue) is seeded with an overnight culture of strain 3AMC, and strains to be tested for killer activity are patched onto the solid media. 11A is a known killer strain, and 2A is a known sensitive strain. Strain 3AM displays an identical response to killer 11A, with a clear zone and methylene blue stained border around the patch of growth.

Figure 4 illustrates the results of transverse alternating field electrophoresis of chromosomes isolated from strains 3AM and 3AMC.

Figure 5 illustrates the results of electrophoresis of dsR A species from killer strains 11A and 3AM, and sensitive strains 2A and 3AMC.

Figure 6 shows yeast growth (Panel A) and sugar utilisation (Panel B) curves of strains 3AM (•) and 3AMC (o) .

Figure 7 shows growth curves of control single monoculture ferments and mixed culture ferments. Panel A: Growth curves of control single monoculture ferments. Symbols: ft 3AM; O 3AMC; Q 5A. Panel B: Growth curves of mixed culture ferments. Symbols: • 3AM and 5A at an inoculum ratio of 2:1; o 3AMC and 5A at an inoculum ratio of 2:1; 3AM and 5A at an inoculum ratio of 1: 1; 3AMC and 5A at an inoculum ratio of 1:1.

Figure 8 shows growth curves of each strain in mixed culture ferments expressed as colony forming units (cfus)/ml. Panel A: Mixed ferment of 3AM(β) and 5A (α) at an inoculum ratio of 1:1. Panel B: Mixed ferment of 3AMC (o) and 5A (α) at an inoculum ratio of 1:1. Panel C: Mixed ferment of 3AM (•) and 5A (α) at an inoculum ratio of 2:1. Panel D: Mixed ferment of 3AMC (o) and 5A (α) at an inoculum ratio of 2:1.

Figure 9 shows proportions of each strain in mixed ferments expressed as percentage of the total yeast population. Panel A: Mixed ferment of 3AM (•) and 5A (a.) at an inoculum ratio of 1:1. Panel B: Mixed ferment of 3AMC (σ) and 5A (o) at an inoculum ratio of 1:1. Panel C: Mixed ferment of 3AM (#) and 5A (o) at an inoculum ratio of 2:1. Panel D: Mixed ferment of 3AMC (o) and 5A (α) at an inoculum ratio of 2:1.

Figure 10 shows the time course in the proportion of killer strain 3AM in the total population of a mixed culture ferment with strain 5A for different inoculum ratios. Symbols: ratio 3AM to 5A • 2:1; O 1:1; • 1:2; en 1:4. DETAILED DESCRIPTION OF THE INVENTION

Preferred embodiments are described below. EXAMPLE I: A CONSTRUCTION OF MARKER SYSTEM FOR WINE YEAST

MATERIALS AND METHODS

Strains and media: Yeast strain AWRI 3A (also known as AWRI 796) was obtained from The Australian Wine Research Institute. E. coli strain DH5a was used for the propagation of all recombinant plasmids. Yeast growth media was YPD [1% yeast extract (Difco) , 2% bacto-peptone (Difco) and 2% glucose] or SD [0.67% Bacto yeast nitrogen base without amino acids (Difco) and 2% glucose] . Plasmid pWX509 was obtained from G.P. Casey, Anheuser-Busch Co., St. Louis. Plasmid pKLG4 was obtained from V. Walbot, Stanford university. (Plasmid pKLG4 is also commercially available.) GUS-vβctor construction: Transformation of DH5a was performed according to the method of Hanahan (30). Rapid plasmid isolation was as described by Ish-Horowicz and Burke (31). DNA fragments were isolated with a GENE CLEAN (BIO 101) kit following supplier's instructions.

Dephosphorylation of fragment ends was achieved by incubation of DNA with 1 unit of calf-intestinal alkaline phosphatase (Boehringer-Mannheim) for 30 minutes at 37^*C in buffer as recommended by supplier. DNA ligations were carried out using T4 DNA ligase (Boehringer-Mannheim) in the recommended buffer.

Yeast transformation and analysis: Yeast cells were transformed using the procedure of Ito et al . (32). Transformants were selected on SD media containing sulfometuron methyl (lOwug/ml) . Total DNA was isolated from transformants according to the method of Davis et al . (33). Southern blotting and hybridizations were carried out by the procedure of Southern (34) with minor modifications. The membrane used was Hybond N÷(Amersham) , and DNA fixation was achieved by exposing the membrane to UV light for 5 mins. Hybridizations were performed at 42"C in solutions consisting of 4% polyethylene glyσol 4000, 2 x SSPE, 1% SDS, 50% formamide, 0.5% blotto and carrier DNA (0.5 mg/ml final concentration) . DNA probes were prepared with an oligolabelling kit from Amersham.

Pulsed field gel electrophoresis: Transverse alternating field electrophoresis (TAFE) was carried out using a Geneline (Beckman) unit. Yeast chromosomes were prepared in agarose plugs according to supplier's recommendations. Electrophoresis was carried out at a constant current of 150 mA; pulse times were 60 seconds for 18 hours, then 35 seconds for 6 hours.

Fermentation trials: Rhine riesling must containing 200g/l reducing sugars was used. Fermentations were carried out in 250ml conical flasks fitted with airlocks. The juice was sparged with N_ gas prior to inoculation, and fermentations were carried out at 18"C with agitation (approximately 100 r.p.m.) under anaerobic conditions. Samples were removed anaerobically during fermentation by needle and syringe through ports covered with rubber septums.

Samples were analysed for sugar level by refractometer readings, and yeast growth by measuring spectrophotometric absorbancy at 650 nm. Residual sugars and acetic acid concentrations were determined with appropriate kits from Boehringer-Mannheim. Measurement of sulfur dioxide in wine was as according to Rankine and Pocock (35). The alcohol content was determined by near infra-red reflectance (36).

GUS assays:

Enzyme assays: Yeast cells were grown to late log phase, then 1ml was pelleted by centrifugation, and the cells resuspended in GUS extraction buffer (50mM NaPO. pH7.0, lOmM beta-mercaptoethanol, lOmM Na-EDTA, 0.1% sarcosyl, 0.1% Triton X-100) . Glass beads (Sigma, 1000-1050 microns) were then added to approximately half the volume, and the suspension was vortexed for 10 mins at 1000 r.p.m. After centrifugation, the supernatant was removed and used as the cell extract.

1-50 l extract was added to 0.5 ml assay buffer (ImM MUG in extraction buffer) and left to incubate at 37'C. At various time intervals, 100 A.1 samples were removed from the assay mix and added to 900ul stop buffer (0.2M Na₂C0-). Solutions were then assayed in a spectrophotofluorometer with xenon lamp (Aminco SPF-125™), excitation 365 nm, emission 455 nm. Standard solutions of 4-methylumbelliferone (MU) (Sigma) in the range of 100 nM to 1^/ULM were prepared for reference values. 1JJ. M MU corresponded to 100 relative units.

Plate assays: Yeast colonies were grown on solid YPD media containing 50-100 ;j.g/ml x-gluc. After approximately 36 hours growth, a solution containing 0.1M NaPO. pH7.0, 1% sarcosyl, 50j g/ml x-gluc and 0.7% agarose was poured as a thin overlay on the plate and allowed to set. After 4-5 hours incubation at 37"C, a blue precipitate could be detected in the transformed colonies. RESULTS GUS-vector Construct

The efficiency of the E. coli B -glucuronidase gene as a marker gene in yeasts is dependent upon maintenance of the gene in a growing population and sufficient gene expression. Expression of the GUS coding region was achieved by use of the yeast alcohol dehydrogenase (ADC1) promoter and terminator sequences. The GUS gene was isolated as a Hindlll fragment from the plasmid pKLG4. This was ligated into the Hindlll site of vector AAH5 (37). A clone with GUS coding region in the correct orientation with respect to the ADC1 promoter was identified (plasmid pAW219). Digestion was carried out with BamHI. This digestion resulted in excision of the GUS gene flanked with ADC1 expression signals (see Figure 1) .

Plasmid pWX509 (38) has been shown to integrate into chromosome XIII of a wine yeast strain and be maintained under fermentation conditions without adverse effects on yeast performance. The GUS plus ADC1 signals cassette was cloned into the BamHI site of vector pWX509 - giving rise to plasmid pAW220 (Figure 2 ) .

TRANSFORMATION AND FERMENTATION TRIALS The SMR1-410 gene on plasmid pAW220 is almost identical in sequence to the ilv 2 gene on the host chromosome - a single base point mutation causes the herbicide resistant phenotype. Therefore, sufficient homology exists between the two sequences to target integration to the ilv 2 gene via homologous recombination with SMR1-410. Recombination is enhanced by digesting plasmid pAW220 with PvuII prior to transformation, giving rise to a linear molecule with SMR1- 410 sequences at either end. The DNA ends are highly recombinogenic, and integration is most likely to occur at the PvuII site in the ilv 2 gene. The result of this event will be two ilv 2 genes flanking the GUS cassette, one of which will contain the SMR1-410 mutation conferring herbicide resistance upon the transformed cell. PvuII digested plasmid pAW220 was introduced into wine yeast strain AWRI 3A by the method of Ito et al (9). Transformants were selected for resistance to the herbicide sulfometuron methyl (

. A transformed colony (designated AWRI 3AM) was then screened for the presence of the GUS construct. Total DNA was isolated and digested with PvuII, thereby releasing the GUS-vector construct intact from the chromosomal DNA. A Southern hybridization was then performed using the Hindlll fragment from plasmid pKLG4 to probe for the GUS sequence. A band of approximately 9kb was evident in the transformed strain, indicating the presence of the GUS construct.

The chromosomal location of the GUS construct in the transformed strain was determined. Intact chromosomes were prepared from strain 3AM, separated by pulsed field gel electrophoresis on a TAFE unit and screened by Southern hybridization for the GUS sequence. Results of this analysis showed that the GUS construct has indeed integrated specifically into the chromosome (XIII) on which the ilv 2 gene is located.

The 3AM strain was used in a fermentation trial to monitor the effects of transformation on the yeast oenological properties. Three different transformants were isolated and used to inoculate separate starter cultures.

Three colonies of control, un-transformed AWRI 3A yeast were also inoculated into starter cultures. Each of these six cultures was inoculated in duplicate into flasks of Rhine g

Riesling grape juice at a concentration of 4 x 10 cells/ml. Fermentations were carried out under anaerobic conditions at 18"C. Samples were taken at regular intervals and assayed for yeast growth (by measuring optical density at 610 nm) and sugar content (by refractive index) . Refractive indices were averaged for both control ferments and for transformant ferments, and the two resulting sugar utilization curves were plotted. There are no significant differences between the fermentation curves of the control and transformed strains. On completion of fermentation, pH, residual sugars, sulfur dioxide, alcohol and acetic acid concentrations were measured for all twelve samples. Averages were calculated for each parameter (see Table 1 below) . Again, no significant differences were evident between the two strains. Finally, in order to test the stability of the GUS construct in the total population, a sample of yeast was recovered from the finished 3AM ferments and assayed for GUS activity. Approximately 500 colonies were analysed by the x-gluc plate method described below; all of these colonies were positive for GUS activity. This result indicates that the introduced construct is maintained in a growing yeast population under fermentation stress, and in the absence of selection by sulfometuron methyl.

# Variance ratio between the two strains

* F Prob. indicates the probability of the associated variance ratio

AWR1-3A - Control unmodified yeast AWR1-3AM - Modified yeast GUS Assays

^-glucuronidase cleaves the substrate x-gluc to release an indoxyl derivative which, upon oxidation, gives rise to an indigo blue dye. GUS activity can be detected by the presence of a blue precipitate. Initial incubations of transformed yeast cells in growth media containing x-gluc at 37'C gave no indication of GUS activity over a three hour period. One possible explanation for this observation is that the yeast cells were unable to take up the x-gluc substrate. Transformed cells were then vortexed for two minutes in the presence of glass beads (to break the permeability barrier) prior to incubation with x-gluc. This treatment resulted in the formation of a blue precipitate within one hour incubation at 37"C. These results indicated that the GUS construct had been successfully introduced into the wine yeast strain, and was being expressed. However, the x-gluc was not taken up by the cells nor could it diffuse into the cells.

A method was devised with these results in mind to assay for GUS activity in yeast colonies. Yeast cells were plated out on YPD media (containing x-gluc) at an appropriate dilution to obtain single, well spaced colonies. After incubation at 30^*C for approximately 36 hours, a molten solution of 0.7% agarose and 1% sarcosyl was poured over the colonies and allowed to set. After incubation at 37^*C for 4- 5 hours, a blue precipitate could be detected in the transformed colonies. This method enables the proportion of marked strain in a total yeast population to be determined simply by calculating the proportion of blue to white colonies.

For more rapid analyses of GUS activity, a sample of yeast cells can be incubated with x-gluc in the presence of 1% sarcosyl until a deep blue precipitate is formed. The proportion of blue to white colonies can be determined under a microscope.

An alternative assay for GUS activity can be performed using the MUG substrate; ^-glucuronidase cleaves this compound to give rise to a fluorescent product. This assay is very sensitive, and was used to determine the GUS enzyme activity.

EXAMPLE II: AN APPLICATION OF MARKER SYSTEM

BACKGROUND

Killer activity in yeasts was first reported in strains of Saccharomyces cerevisiae in 1963 by Bevan and Makower (7). Killer yeasts secrete polypeptide toxins which kill sensitive strains of the same genus and less frequently, strains of different genera (8, 9). Previous studies indicate that the toxin of Saccharomyces is a protein which binds to a receptor on the cell wall of the sensitive yeast, disrupting the electrochemical gradient across the cell membrane and hence the intracellular ionic balance (10, 11).

Production of the toxin and immunity to it are determined by a cytoplasmically inherited double stranded (ds) RNA plasmid, otherwise known as the M-genome (12), which is found only in cells containing an additional dsRNA species designated the L-genome. Both types of dsRNA exist in virus¬ like particles and require a protein encoded by the L-dsRNA for encapsidation (12, 13).

Based upon properties of the toxin, killer yeasts have been classified into eleven groups (Kl through Kll) (14, 15). Those unique to Saccharomyces fall into the first three (Kl, K2 and K3). The Saccharomyces toxin is reversibly inactivated at low pH (2.0) and the irreversibly inactivated at pH in excess of 5.0 (15). More specifically, the biological activity of Kl is optimal between pH 4.6 and 4.8 while K2 shows optimal activity between 4.2 and 4.7 (16). Compared to Kl, the K2 toxin is stable over a wider pH range (2.8 to 4.8) (17) and is therefore more relevant in wine fermentation. Killer activity has been detected in yeasts isolated from established vineyards and wineries in various regions of the world, including Europe and Russia, South Africa and Australia. This widespread occurrence has prompted interest in the oenological significance of killer wine yeasts. In theory, selected killer yeasts could be used as the inoculated strain to suppress growth of undesirable wild strains of Saccharomyces cerevisiae during grape juice fermentation. In addition, as killer interactions have been reported to occur between yeasts of different genera (18, 19), the possibility exists to genetically engineer broad spectrum killer strains of yeasts such as Saccharomyces cerevisiae (20).

Previous studies have been conducted to assess the efficiency of killer toxin on sensitive yeast strains. However, reports have been contradictory on the expression of killer activity under fermentation conditions (4, 21, 22). Attempts to determine the population kinetics of killer and sensitive strains during wine fermentation have been restricted because of the difficulty involved in identifying two types of strains when grown in mixed cultures. Approaches used to date include i) choice of killer and sensitive strains that can be distinguished by their growth rates (23) or production of hydrogen sulfide (24); ii) use of auxotrophic and respiratory deficient mutants of killer strains and appropriate plating conditions under which they can be identified (25, 26, 27); iii) use of killer and sensitive strains which can be distinguished by differences in colony morphology (28); and iv) assaying colonies directly for killer activity (29) . All of these methods are limited by the fact that the assays involved are laborious and time- consuming, or that only killer strains with specific characteristics can be studied.

MATERIALS AND METHODS

Strains and media: Sensitive Saccharomyces cerevisiae strains AWRI 5A (also known as AWRI 138), AWRI 2A (AWRI 729), and killer strain AWRI 11A (AWRI 92F) were obtained from the Australian Wine Research Institute. Generation of the marked killer strain 3AM (AWRI 796) has been previously described (39). Yeast growth media was YPD [1% yeast extract (Difco), 2% bacto-peptone (Difco) and 2% glucose] .

Curing of Killer strain 3AM: A culture of strain 3AM was grown overnight in YPD at 28^*C. Serial dilutions were made in 0.9% NaCl and 0.1 ml aliquots (containing approximately 100 cells) were spread on YPD plates and incubated at 37'C. After 48 hours incubation, single colonies were selected at random and assayed for killer activity as described below.

Assay for cured strain: YPD (containing 1% agar) was sterilized by autoclaving at 120"C for 20 mins. After cooling to 49'C, the medium was buffered to pH 4.2 with 0.05M tartrate buffer. Methylene blue (to 0.003% w/v) and killer

5 sensitive strain 5A (to 10 cells/ml) were added to the medium prior to pouring the plates. Colonies isolated after heat treatment were then patched onto these assay plates and incubated at 18'C for approximately 72 hours. Curing was recognised by the absence of growth inhibition (clear zones) and lack of blue stained cells around the patched colony. dsRNA isolation: The dsRNA extraction procedure was essentially that described by Fried and Fink (40) . Samples of RNA were analysed by electrophoresis on 1.5% agarose slab gels at a constant current of 100 mA. Gels were stained with ethidium bromide and photographed on a short wave UV light box.

Fermentation trials: Starter cultures were prepared by inoculating 10 ml of YPD medium contained in a conical flask with a loopful of yeast and incubated with vigorous aeration at 28^*C. After 24 hours, the cell density was determined by microscopic counts. Samples were used to inoculate Rhine Riesling must (200 ml) to a density of 5 x 10 cells/ml. The must contained 220 g/1 reducing sugars and had a pH of 3.1. Fermentations were carried out in 250 ml conical flasks fitted with airlocks. The juice was sterilized by membrane filtration prior to inoculation, and fermentations were carried out at 18^*C with agitation (approximately 100 r.p.m.). Samples were removed anaerobically during fermentation by needle and syringe through ports covered with rubber septurns.

Samples were analysed for the progress of fermentation by refracto eter readings, and yeast growth by measuring spectrophotometric absorbancy at 650nm. For analysis of the proportion of marked strain in the yeast population, serial dilutions were made of the samples in sterile 0.9% NaCl and 0.1 ml aliquots were plated (containing 200-500 cells) on YPD media. The plates were then assayed as described below.

GUS plate assays: Yeast colonies were grown on solid YPD media for approximately 36 hours at 28'C. A solution containing 0.1M Na₂HPO., pH 7.0, 1% sarcosyl, x-gluc (100-150 ug/ml) and 0.7% agarose was then poured as a thin overlay on the plate and allowed to set. After 4-6 hours incubation at 37^*C, a blue precipitate could be detected in the marked colonies.

RESULTS

Curing of Strain AWRI 3AM

In order to specifically analyse the effect of killer toxin in fermentations, an experiment was designed to compare two isogenic strains which differ only in the presence of the M-dsRNA genome and therefore in their ability to produce killer toxin.

Killer strain 3AM had previously been marked with the Escherichia coli GUS gene (39) . This system allows the marked strain to be readily identified in a mixed population by a simple plate assay which results in the formation of a blue precipitate in marked colonies. Strain 3AM was cured of its M-dsRNA plasmid by heat treatment (41), the cured or sensitive colonies being identified by killer plate assays. A zone of inhibition clearly evident around strain 3AM was absent around 3AMC, indicating that strain 3AMC is not producing killer toxin (Figure 3).

In order to verify that the isolated strain 3AMC is a genuine derivative of strain 3AM, both strains were karyotyped by pulsed field gel electrophoresis. Total chromosomes were isolated and electrophoresed on a Transverse Alternating Field Electrophoresis (TAFE) system (Figure 4) . An identical eleσtrophoretic pattern was obtained for both strains, thus illustrating a common genetic background. As strain 3AMC is derived from 3AM, it inherits the GUS gene and therefore is also a marked strain.

Finally, dsRNA species were isolated from strains 3AM and 3AMC and analysed by standard electrophoresis techniques (Figure 5). A band representing the M-dsRNA genome is present in strain 3AM, and absent in strain 3AMC.

Fermentation trials were then performed on strains 3AM and 3AMC to determine the effect of the curing procedure on yeast growth and fermentation rates. Starter cultures of each strain were inoculated in triplicate into flasks of Rhine g Riesling grape juice at a concentration of 5 x 10 cell/ml.

Samples were taken at regular intervals and assayed for yeast growth and progress of fermentation. The average readings for each strain were plotted over time (Figure 6). There were no significant differences in the growth or fermentation rates between strains 3AM and 3AMC.

ANALYSIS OF KILLER ACTIVITY DURING FERMENTATION

Strains 3AM and 3AMC were analysed for killer activity in Rhine Riesling juice by co-inoculating each strain with the sensitive Saccharomyces strain 5A. Control ferments of each strain (3AM, 3AMC and 5A) as pure inoculums were also performed. Each ferment was conducted in duplicate at 18"C with gentle agitation under anaerobic conditions. GUS plate assays were then performed to identify the marked strain (3AM or 3AMC) . Colonies of the marked strain turn a deep blue colour as a result of this assay, allowing simple identification. GUS plate assays were also performed on the control ferments to confirm the validity of the assay. Plate assays on the control 5A ferment were consistently negative, highlighting the absence of background GUS activity in natural yeast cells. However, control 3AM and 3AMC ferments gave values of between 99 - 100% for the marked strain count. In other words, occasionally colonies of a marked strain did not turn blue in response to the GUS assay. The frequency of this occurrence was always less than 1% of the total plate count, and did not increase over time. This background reversion frequency was taken into account throughout the analysis.

The following mixed culture ferments were carried out: i) 3AM and 5A at an inoculum ratio of 1:1 respectively; ii) 3AMC and 5A at an inoculum ratio of 1:1 respectively; iii) 3AM and 5A at an inoculum ratio of 2:1 respectively; and iv) 3AMC and 5A at an inoculum ratio of 2:1 respectively. These mixed ferments exhibited normal growth kinetics, as did the three control ferments (Figure 7) .

The time course of growth (colony forming units per ml) of each strain in the mixed culture ferments was plotted in Figure 8. At inoculum ratios of 1:1, there was a notable increase in the proportion of killer strain 3AM, whereas the cured strain 3AMC failed to exert any dominance over the sensitive strain under otherwise identical conditions. Statistical analysis was used to test the null hypothesis that the ratio of killer: sensitive cells remains 1:1 throughout the ferment. A "goodness of fit" test (normal test) rejected the null hypothesis, with p-value << 0.001. However, identical analysis of the cured : sensitive strain ferment supported the null hypothesis that the ratio of the two strains remains at 1:1 throughout the ferment.

With an increased proportion of strains 3AM and 3AMC in the inoculum (ratio 2:1), the dominating effect of strain 3AM was more pronounced, whereas strain 3AMC again showed no apparent change in proportion over time. The dominance of strain 3AM in mixed culture ferments was illustrated more clearly when the percentage of each strain was plotted over the time of the ferment (Figure 9) . For an inoculum ratio of 1:1, 3AM increased to approximately 80% after 3 days, but for a higher inoculum ratio of 2:1, 3AM eventually accounted for 97% of the total yeast population. It is important to note that the strain 5A persisted, albeit at low levels, throughout the ferments . This contrasts with previous reports of killer activity, where the sensitive strain was not detectable by the end of the ferment (28, 29).

EFFECT OF INOCULUM RATIO ON KILLER ACTIVITY Experiments were conducted to determine the lowest inoculum ratio of killer to sensitive cells at which significant killer activity could be observed. Mixed ferments of strain 3AM and 5A at inoculation ratios of 1:2 and 1: 4 respectively were carried out under conditions described above.

At various times during the ferment, samples were analysed for the proportion of strain 3AM in the total population (Figure 10) . An increase in proportion of strain 3AM was evident when it was inoculated at greater than 50% of the total population. However, when it was less than 50% at the time of inoculation, there was no change in proportion.

DISCUSSION Previous studies have indicated 100% stability of the GUS marker gene in strain 3A throughout fermentation (39). However, analysis of a larger sample of colonies in these experiments has revealed an instability of the construct. This instability was detected in the control ferments which were inoculated with monocultures of either of the marked strains (3AM or 3AMC) . Samples from these ferments gave rise to colonies which responded negatively to the GUS plate assay at a frequency of less than 1% of the total plate count. Occasionally a colony which was sectored in its response to the assay was detected, suggesting either excision of the gene by homologous recombination (45) or loss of the gene after mitotic crossing-over (46). The frequency of instability did not increase over time during fermentation, and could be directly quantified in the control 3AM and 3AMC ferments.

This marking system has enabled a direct comparison to be made between the inoculation efficiency of a killer strain (3AM) and a cured derivative of the same strain (3AMC) in fermenting grape juice. An eleσtrophoretic karyotype of the killer and cured strains confirmed a common genetic constitution, where analysis of dsRNA species revealed the loss of the M-dsRNA genome in the cured strain. No differences were detected in the growth rates or fermentation curves of strains 3AM and 3AMC. Therefore, by comparing strain 3AM and 3AMC, differences in properties can be attributed directly to the presence of the M-dsRNA plasmid and hence to the production of the K2 killer toxin.

Under identical fermentation conditions, the cured strain 3AMC remained at 50% of the total population while the killer strain increased to 80%. The ability of strain 3AM to dominate 5A during fermentation is likely to be due to the production of killer toxin by strain 3AM and not to a difference in respective growth rates favouring the killer strain. We can conclude, therefore, that the killer toxin has displayed significant activity under these fermentation conditions. This result is of particular interest to the oenologist since the K2 toxin produced by strain 3A is reported to show maximum activity at pH 4.2 (17), which is 0.5 to 1 pH unit higher than generally found in grape musts.

In cases where killer activity has been reported in fermenting grape juice, a discrepancy exists as to whether effective killing action occurs when the proportion of killer cells is less than 50% of a mixed culture ferment. Heard and Fleet (28) did not observe killer action when the ratio of killer to sensitive cells was approximately 1:7, whereas others have reported killer activity with killer to sensitive cell ratios of 1:10 and lower (23, 25, 26).

Our results showed that an increase in ratio of killer to sensitive cells to approximately 2:1 resulted in a pronounced dominance of the ferment by strain 3AM to 97% of the total mixed population by the end of the ferment. However, with killer to sensitive cell ratios of 1:2 or 1:4, no effective killer action was evident. It is possible that differences in either composition of medium, fermentation conditions or strain sensitivity may account for discrepancies in reports of killer toxin efficiency.

The relevance of killer strains in wine making has been the focus of attention in countries where selected yeast cultures are inoculated into musts to induce fermentation. This focus has intensified since the observations that yeasts which are naturally present in the must also play significant roles in supposedly "pure" culture fermentations (22, 42). These natural yeasts include species from the genera Kloeckera, Candida, Hansenula and Saccharomyces. Killer Saccharomyces wine yeast strains may be effective in suppressing natural Saccharomyces yeasts during fermentation, and the possibility exists to engineer broad range killer yeasts to control strains from other genera. For these reasons, further study is needed to determine appropriate fermentation conditions for effective killer activity.

The GUS marking system provides a method which allows a broad range of killer strains to be rapidly and unequivocally identified in a mixed culture. This system can be employed to gain a better understanding of killer activity during fermentation.

DISCUSSION AND CONCLUSIONS

The GUS-vector construct described in this specification can be introduced to a range of yeast strains by transformation procedures, e.g. using the SMR1-410 gene as a dominant selection marker. Results show that the construct can be integrated into a specific site in the yeast genome without disrupting essential functions or affecting the fermentation performance of a wine yeast strain. Once integrated into the genome, the construct is maintained in a stable manner throughout the fermentation. Assaying for the GUS marker can be achieved by fluorimetry, spectrophotometry or by agar plate assay. Although natural transport of x-gluc or MUG substrates did not occur across yeast cell membranes in the time course of experiments described here, this problem was overcome by inducing artificial permeation in assay procedures. Methods have been described which allow the proportion of marked strain in a total yeast population to be determined.

The applications of this marker system will be in monitoring inoculated strains under different fermentation conditions. Recent investigations by Heard and Fleet (42, 43, 44) made with both inoculated and uninoculated grape juices under a range of fermentation conditions suggest that Saccharomyces strains are not necessarily the dominant organism during vinification. These studies, along with the lack of knowledge regarding incidence and importance of wild strains of Saccharomyces in fermenting grape juice, call for detailed monitoring of inoculated strains under varying oenological environments. The GUS system described in this specification will enable a wide range of yeast strains to be marked, providing the means for unequivocal identification and monitoring during fermentation.

REFERENCES

1. Rankine, B.C. The importance of yeasts in determining composition and quality of wines. Vitis 2:22-49 (1968)

2. Vezinhet, F. and Lacroix, S. Marquage genetique de levures: Outil de controle des fermentations en souche pure. Bull, de l'O.I.V. 643-644 : 759-773 (1984)

3. Loiseau, G., Vezinhet, F., Valade, M. , Vertes, A., Cuinier, C. and Delteil, D. Controle de 1'efficacite du levurage par la mise en oeuvre de souches de levures oenologiques marquees. Rev. Fr. Oenol. 106: 29-36 (1987) 4. Delteil, D. and Aizac, T. Comparison of yeast inoculation techniques by the use of a 'marked' yeast strain. Aust. and N.Z. Wine Industry Journal, November 1988: 53-56

5. Jeffersen, R.A. , Burgess, S.M., and Hirsch, D. β-glucuronidase from Escherichia coli as a gene fusion marker. Proc. Natl. Acad. Sci. USA .83:8447-8451 (1986)

6. Jeffersen, R.A. , Kavanagh, T.A. , and Beven, M.W. GUS fusions: β-Glucuronidase as a sensitive and versatile gene marker in higher plants. EMBO J. 6.:3901-3907 (1987)

7. Bevan, E. A., and M. Makower. The physiological basis of the killer character in yeasts. In: Genetics Today, Procedings of the 11th International Congress of Genetics, 1 (1963). S. J. Gert (ed.). pp 202-203.

8. Philliskirk, G. , and T. W. Young. The occurrence of killer character in yeasts of various genera. Antonie van Leeuwenhoek 4JL: 147-151 (1975).

9. Tipper, D. J. and K. A. Bostian. Double-stranded ribonucleic acid killer systems in yeasts. Microbiol. Rev. : 125-156 (1984).

10. De le Pena, PI, Barros, F. , Gascon, S., Lazo, P. S. and Ramos, S. Effect of yeast killer toxin on sensitive cells of Saccharomyces cerevisiae. J. Biol. Chem.

15_6: 10420-25 (1981).

11. Skipper, N. and H. Bussey. Mode of action of yeast toxins: energy requirement for Saccharomyces cerevisiae killer toxin. J. Bacteriol. 119: 668-677 (1977). 12. Bostian, K. A., Hopper, J. E., Rogers, D. J. and Tipper, D. J. Translational analysis of the killer-associated virus-like particle dsRNA genome of Saccharomyces cerevisiae: M-dsRNA encodes toxin. Cell _L9_: 403-414. (1980) .

13. Harris, M. S. Virus-like particles and double-stranded RNA from killer and non-killer strains of Saccharomyces cerevisiae. Microbiology !: 161-174 (1978).

14. Naumov, G. I. and T. I. Naumova. Comparative genetics of yeast communication. XIII. Comparative study of killer strains of Saccharomyces from different collections. Genetika 1:140-145 (1973).

15. Young, T. W. and Yagiu, M. A comparison of the killer character in different yeasts and its classification. Antonie van Leeuwenhoek J. Microbiol. 44: 59-77 (1978).

16. Shimazu, K. , Adachi, T. , Kitano, K. , Shimazaki, T. , Totsuku, A., Hara, S., and H. H. Dittrich. 1985. Killer properties of wine yeasts and characterisation of killer wine yeasts. J. Ferment. Technol. 63: 421-429.

17. Rogers, D., and E. A. Bevan. 1978. Group classification of killer yeasts based on cross reactions between strains of different species and origin. J. Gen. Microbiol. 105: 199-202.

18. Radler, F. , Pfeiffer, P., and M. Dennert. 1985. Killer toxins in new isolates of the yeasts Hanseniaspora uvar m and Pichia kluyveri . FEMS Microbial Lett. 29: 269-272. 19. Rosini, G. Interaction between killer strains of Hansenula anomala var. anomala and Saccharomyces cerevisiae yeast species. Can. J. Microbiol. 31:300-2 (1985) .

20. Boone, C. , Sdicu, A. M. , Wagner, J., Degre, R. , Sanches, C. , and H. Bussey. 1990. Integration of the yeast Kl killer toxin gene into the genome of marked wine yeasts and its effect on vinification. Am. J. Enol. Vitic. 41: 37-42.

21. Cuinier, C. and C. Gros. Enquete sur la repartition des levures killers en France. Vivnes et vins 23.- 25-27 (1983) .

22. Lafon-Lafourcade, S. and P. Ribereau-Gayon. Developments in the microbiology of wine production. Prog. Ind. Microbiol. 19:1-45 (1984).

23. Barre, P. Le mecanisme killer dans la concurrence entre souches de levures. Evaluation et prise en compte. Bull. O.I.V. 641-642: 635-643 (1984).

24. Rosini, G. Effet d'une levure "killer" de Saccharomyces cerevisiae sur une souche de levure sensible de la meme espece non productrice de H2S et selectionee pur la vinification dans un milieu de culture mixte. Bull. O.I.V 648-649: 214-217 91985).

25. Hara, S., Iimura, Y. and F. Otsuku. Breeding of useful killer wine yeasts. Am. J. Enol. Vitic. .12:28-33 (1980).

26. Hara, S., Iimura, Y. , Oyama, H., Kozeki, T. , Kitano, K. and Otsuku, K. Breeding of cryophilic killer wine yeasts. Agric. Biol. Chem. .45.: 1327-1334 (1981). 27. Seki, T. , Choi, E.H. and D. Ryu. Construction of killer wine yeast strains. Appl. Environ. Micro. 49: 1211-1215 (1985) .

28. Heard, G.M. and Fleet, G. H. Occurrence and growth of killer yeasts during wine fermentation. Appl. and Environ. Microbiol. 11:2171-2174 (1987).

29. Longo, E. , Velazquez, J. B., Cansado, J. , Calo, P. and T. G. Villa. Role of killer effect in fermentations conducted by mixed cultures of Saccharomyces cerevisiae. FEMS Microbiol. Letters 21: 331-336 (1990).

30. Hanahan, D. Studies of transformation of Escherichia coli with plasmids. J. Mol. Biol. 166: 557-580 (1983)

31. Ish-Horowicz, D. and Burke, J.F. Rapid and efficient σosmid cloning. Nucleic Acids Res. .9: 2989-2998 (1981)

32. Ito, H. , Fukuda, Y. , Murata, K. and Kimura, A. Transformation of intact yeast cells treated with alkali cations. J. Baσteriol. 153: 163-168 (1983)

33. Davis, R.W. , Thomas, M. , Cameron, J.R., St. John, T.P., Sσherer,S. and Padgett, R.A. Rapid DNA isolations for enzymatic and hybridization analysis, in Methods in Enzy ol., Vol. 65, Gross, L. and Moldave, K. (eds.), Academic Press Inc., New York, pp 404-411 (1980)

34. Southern, E.M. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol.Biol. £8.: 503-517 (1975)

35. Rankine, B.C. and Pocock, K.F. Alkalimetric determination of sulfur dioxide in wine. Aust. Wine Brew. Spirit Rev. 91: 62,64,66 (1972) 36. Anon. Instructions for the use of the Technicon 260 Infra-analyser. Bran & Luebbe Australia.

37. Gustav, A. Expression of genes in yeast using the ADC1 promoter. In Methods in Enzymol., Vol 101,

(eds), Academic Press Inc., New York, pp 192-201.

38. Casey, G.P., Wei Xiao and Rank, G.H. A convenient dominant selection marker for gene transfer in industrial strains of Saccharomyces yeast : SMR1 encoded resistance to the herbicide sulfometuron methyl. J. Inst. Brew. £4: 93-97 (1988)

39. Petering, J. E., Henschke, P. A. and Langridge, P. The Escherichia coli ^-Glucuronidase gene as a marker for Saccharomyces yeast strain identification. Am. J. Enol. Vitic, Al: 6-12 (1991) .

40. Fried, H. M. and G. Fink. Electron microscope heteroduplex analysis of "killer" double-stranded RNA species from yeast. Proc. Natl. Acad. Sci. U.S.A. 69: 2846-2849 (1978).

41. Wickner, R.B. "Killer character" of Saccharomyces cerevisiae: curing by growth at elevated temperatures. J. Bacteriol. 112: 115-126 (1974).

42. Heard, G.M. and Fleet, G.H. Growth of natural yeast flora during the fermentation of inoculated wines. Applied and Environmental Microbiology .50. : 727-728 (1985)

43. Heard, G.M. and Fleet, G.H. Occurrence and growth of yeast species during the fermentation of some Australian wines. Food Technology in Australia 18, : 22-25 (1986) 44. Heard, G.M. and Fleet, G.H. The effects of temperature and pH on the growth of yeast species during the fermentation of grape juice. Journal of Applied Bacteriology 31 : 23-28 (1988)

45. Struhl, K. , Stinchcomb, D.T., Scherer, S. and R.W. Davis. High-frequency transformation of yeast: Autonomous replication of hybrid DNA molecules. Proc. Natl. Acad. Sci. USA 13: 1035-1039 (1979).

46. Roeder, G.S. and S.E. Stewart. Mitotic recombination in yeast. Trends in Genet. 4.: 263-267 (1988).

Claims

The claims defining the invention are as follows:

1. A DNA sequence comprising a ^-glucuronidase coding region operably linked to yeast promoter and terminator sequences.

2. A DNA sequence according to claim 1, wherein said ^-glucuronidase coding region is derived from Escherichia coli .

3. A DNA sequence according to claim 1 or claim 2, wherein said promoter and terminator sequences are alcohol dehydrogenase promoter and terminator sequences .

4. A DNA sequence according to any one of claims 1 to 3, wherein said promoter and terminator sequences are derived from Saccharomyces cerevisiae.

5. A plasmid comprising a DNA sequence according to any one of claims 1 to 4.

6. Plasmid pAW 220.

7. A method of producing a plasmid according to claim 5 or claim 6, comprising ligating yeast promoter and terminator sequences to a -glucuronidase gene and inserting the resultant construct into a plasmid.

8. A method of integrating a β-glucuronidase coding region into a yeast strain comprising:

(a) ligating yeast promoter and terminator sequences to a ^-glucuronidase gene,

(b) inserting the resultant construct into a carrier plasmid comprising a DNA sequence which is substantially identical to a DNA sequence on the host yeast chromosome, and

(c) utilising the recombinant plasmid from step (b) to integrate the ^-glucuronidase gene, together with yeast promoter and terminator sequences, into the host yeast chromosome.

9. A method according to claim 8, wherein the carrier plasmid comprises the SMR1-410 gene.

10. A method according to claim 8 or claim 9, wherein the carrier plasmid is pWX509.

11. A method according to any one of claims 8 to 10, wherein the yeast strain is a killer yeast strain.

12. An assay for β-glucuronidase activity comprising:

(a) adding a substrate for the glucuronidase gene to a sample of a yeast strain;

(b) inducing permeation in the cell membrane of the yeast strain;

(c) incubating the yeast strain for sufficient time to form detectable quantities of a product of the ^-glucuronidase enzyme reaction; and

(d) testing for said product.

13. An assay according to claim 12, wherein said substrate is 5-bromo-4-chloro-3-indolyl glucuronide or 4- ethyl umbelliferyl glucuronide.

14. An assay according to claim 12 or claim 13, wherein said yeast strain comprises a DNA sequence according to any one of claims 1 to 4.

15. A DNA sequence or plasmid according to any one of claims 1 to 6, substantially as described herein.

16. A method according to any one of claims 7 to 11, substantially as described herein.

17. An assay according to any one of claims 12 to 14, substantially as described herein.