ARCHIVED - Novel Food Information

Wines derived from the genetically modified wine yeast Saccharomyces cerevisiae ML01

Health Canada has notified Dr. Hennie van Vuuren of the University of British Columbia that it has no objection to the food use of a genetically modified wine yeast, Saccharomyces cerevisiae ML01, for use in winemaking in Canada, and of wines derived from the use of this yeast. The Department conducted a comprehensive assessment of wine yeast ML01 according to its Guidelines for the Safety Assessment of Novel Foods. These Guidelines are based upon internationally accepted principles for establishing the safety of foods with novel traits.

Background:

The following provides a summary of the notification from Dr. van Vuuren and the evaluation by Heath Canada and contains no confidential business information.

1. Introduction

Dr. Hennie van Vuuren developed Saccharomyces cerevisiae ML01 using polyethylene glycol (PEG) transformation to introduce a DNA expression cassette into Saccharomyces cerevisiae var. bayanus S92 (Lesaffre Yeast Corporation), via homologous recombination

The purpose of the genetically modified malolactic wine yeast ML01 is to remove, via malolactic fermentation, the malic acid from wines, without the use of starter cultures. This has been achieved by introducing two genes into the yeast's genome, mae 1, which codes for a proton symport that permits the transport of malate, malonate and succinate into the yeast cell and mle A, coding for the malolactic enzyme that converts L-malate to L-lactate. By not using starter cultures, biogenic amines, which can have undesirable health effects, will not be formed.

The assessment conducted by Food Directorate evaluators determined how ML-01 was developed; how it was manufactured; determined the composition of wines made using this yeast; and the potential for the presence of any toxicants, anti-nutrients, allergens, chemical or microbiological contaminants in such wines.

The Food Directorate has a legislated responsibility for pre-market assessment of novel foods and novel food ingredients as detailed in Division 28 of part B of the Food and Drug Regulations ("novel foods"). Food use of ML01 is considered a novel food under the following part of the definition of novel foods: "c) a food that is derived from a plant, animal or microorganism that has been genetically modified such that

1. the plant, animal or microorganism exhibits characteristics that were not previously observed in that plant, animal or microorganism."

2. Development of the Modified Yeast

Saccharomyces cerevisiae ML01 was developed using polyethylene glycol (PEG) transformation to introduce a DNA expression cassette into Saccharomyces cerevisiae var. bayanus S92 (Lesaffre Yeast Corporation), via homologous recombination. The introduced sequence consists of the mleA and maeI genes both under the control of the S. cerevisiaePGK1 promoter and terminator sequences necessary for expression in yeast. The malolatic expression cassette carried on plasmid pJH2, including the flanking 3' and 5' regions of the URA3 gene, was fully sequenced and described by the petitioner.

The maeI gene is derived from Schizosaccharomyces pombe 972h- and codes for the protein malate permease, which permits the entry of malate into the wine yeast. S. pombe is an ascomycete yeast found on fruits (apples and grapes) and food (grape juice, palm wine, sugar cane and syrup) which is used for deacidification of wine and is not known to have adverse effects in humans. The mleA gene is derived from Oenococcus oeni Lo8413 and codes for the malolactic enzyme which breaks down the malic acid inside the yeast. O. oeni is commonly used in the wine industry as a starter culture for malate removal, and has a history of safe use. S. cerevisiae AB972 is the source of the PGK1 (3-phosphoglycerate kinase) promoter and terminator sequences used in the construct. S. cerevisiae GC210 is the source of the URA 3 flanking regions used for the construct integration via homologous recombination. S. cerevisiae has a long history of safe use.

The malolactic cassette was cut from pJH2 yielding the fragment used to transform S. cerevisiae and was introduced into the yeast alongside plasmid pUT332, via PEG transformation. Plasmid pUT332 carries a phleomycin resistance gene (Tn5Ble) used in the initial screening step, and ampicillin resistance (Amp^r). Two-thousand phleomycin resistant colonies were selected and screened for the desired phenotype, of which only 4 strains were positive for the new phenotype, and only one was confirmed to have integrated the cassette in the URA3 gene. This strain is referred to as ML01.

Plasmid pUT332 was removed from ML01 by subculturing it over several passages on a non-selective media and the absence of the plasmid was confirmed by probing the total yeast DNA with the Tn5Ble and the backbone/Amp^rprobe.

3. Characterization of the Modified Organism

Southern blot analysis and sequencing of ML01 DNA demonstrated the insertion of a single intact copy of the malolactic cassette at the URA3 locus in the yeast genome. Furthermore, sequencing data and Southern data indicated that no rearrangements of the DNA had occurred, as all internal restriction enzyme sites remained intact producing hybridization fragments of the expected size. Sequencing of the insert showed that some base substitutions (15) had occurred compared to the published sequence. These are mainly the result of artefacts from the construction process (7 out of 15), while the others may be due to changes introduced during amplification of the various gene sources. Only one of these changes caused an amino acid substitution (Asp to Glu) occurring in the mleA protein. This change in the protein was not considered significant and raised no concerns.

The genetic stability of the inserted DNA was evaluated through Southern analysis of ML01 over 100 generations. Using EcoRI and probe, which yielded 3 distinct bands, the petitioner showed that the insert is stable over at least 100 generations.

Expression of both genes was confirmed by Reverse Transcriptase-PCR (RT-PCR). ML01 was grown in synthetic must, and mRNA was extracted from the must at 2 and 6 days after the start of the fermentation and submitted to RT-PCR. Both genes, mae I and mle A, were shown to be expressed during this period. Furthermore, the degradation of malate was followed (the expected phenotype) during the 6 days, and was observed to be fully degraded.

Genomic data was provided to determine the impact of the insertion event on overall gene expression. The parent strain S92 and ML-01 were grown in synthetic must, the total mRNA was extracted and analyzed in quadruplicate. Out of 6200 Open Reading Frame's (ORF's), only 15 presented a more than 2-fold change, compared to S-92 baseline, with a maximum of a 6-fold change. Of these 15, 6 were upregulated and 9 down regulated. Some of the ORFs which were affected do not have known functions, other ORFs are ribosome biosynthesis and heat -shock proteins which are often influenced by intracellular pH (note that as malate enters the ML-01 cell, the pH should decrease). One particular ORF is BTN2 , which hypothetically would code for a protein involved in pH control in yeast cells. It is thought that most of these changes are attributed to decrease in intracellular pH, due to entry of malate into the cell. Overall, the introduction of the malolactic cassette at the URA3 locus did not affect the overall transcriptome of the yeast.

The insertion event's sequence was analysed for hypothetical open reading frames of more than 90 base pairs with a TATA box in 5' direction. Seven ORFs were found ranging from 108 bp to 660 bp in length. Based on the sequences of these ORFs, primers were designed for RT-PCR to determine if these ORFs were being transcribed. From this analysis all ORFs where shown to produce a transcript. It is unlikely that the shorter transcripts (less than 150 bp: ORFs 2,3,5,7) would be translated into a new protein, as short mRNAs are unstable and have short half-lives. The remaining four transcripts, if they were translated, would share high homology with the mae 1 and mle A expressed proteins and therefore were not considered to be of any concern.

4. Product Information

Malolactic wine yeast ML01 differs from its traditional counterparts due to insertion of two genes mleA and maeI under the control of promoter and terminator sequences necessary for expression in yeast. The insertion of these genes in yeast result in the expression of malolactic enzyme and malate permease respectively. The expression of these proteins in the yeast allows for the removal of the malic acid during the production of wine without the use of starter cultures.

5. Dietary Exposure

Due to the manufacturing process itself, very little yeast is left in wines, so exposure to the yeast is minimal. Wine is normally left to sediment to clarify it, with final counts of 1000/10 000 CFUs/ml wine. Wines may be further filtered on diatomaceous earth, where counts are reduced to less than 1000 cells/100 ml, and even further, on cellulose filter (50 cells/100 mL). It is usual practice nowadays to filter wine on 0.45 um filters for white wine, 0.65 um for red wine, which will eliminate all yeast cells. Wine consumption itself is not expected to change.

6. Nutrition

Wine is not consumed for its nutritional value. Therefore, it was concluded that it was not necessary to require compositional data that compare wine made with the novel yeast ML01 with wine made by traditional methods, from a nutritional standpoint.

7. Chemistry

There are Canadian guidelines for ethyl carbamate in wines. For table wines, the guideline is 30 ppb. For fortified wines, the guideline is 100 ppb. The petitioner compared ethyl carbamate precursor levels in three Merlot wines: (1) one produced using the ML01 yeast; (2) one produced using the parental yeast strain S92; and (3) one that underwent alcoholic fermentation using the strain S92 and malo-lactic fermentation using lactic acid bacteria (Oenococcus oeni). Ethyl carbamate precursors were indirectly measured by heating the wine samples at 72 C for 48 hours to accelerate ethyl carbamate formation and then measuring the ethyl carbamate concentration as an indicator of the levels of precursors. The heated wine samples from the ML01 strain and the S92 strain contained 12.0 and 13.0 ppb ethyl carbamate respectively. The wine produced using S92 and O. oeni contained 20.3 ppb ethyl carbamate. These levels are below the guideline for table wines and therefore were not considered to be of any health concern.

The finding of higher levels of ethyl carbamate precursors in the wine that underwent bacterial malo-lactic fermentation is consistent with the fact that such precursors are known to be formed by bacteria during malo-lactic fermentation. The finding that levels were similar in the ML01 and S92 alcoholic fermented wines indicates that wine produced using ML01 should not lead to higher ethyl carbamate levels than found in wines that have been fermented using conventional microorganisms.

There is no Canadian guideline for histamine or other biogenic amines in wines. The formation of biogenic amines in wine is associated with metabolic pathways in lactic acid bacteria, although not all bacteria carry this type of activity. While the petitioner has not analysed for biogenic amines in the wine produced from ML01 yeast, it is asserted that there is "no reason to suspect that the ML01 strain will produce more biogenic amines than the parental S92 strain." Several reasons are given: (1) in the ML01 yeast, the bacterial gene encoding for malolactic enzyme would not have activity that would lead to the formation of biogenic amines in wines; (2) Saccharomyces cerevisiae has not been shown to be involved in the synthesis of biogenic amines in wine. Therefore, the petitioner expects that wines that have undergone malo-lactic fermentation by the ML01 strain will contain lower levels of biogenic amines than wines that have undergone malo-lactic fermentation by lactic acid bacteria.

8. Microbiology

Yeast ML01 will be produced in the same manner as other wine yeast and the petitioner has provided a flow diagram describing the batch fermentation process. As there are no required Food Chemical Codex specifications for wine yeast, the petitioner has provided the Springer Oenologie's microbiological specifications for the product. These specifications state that ML01 contains: (1) no more than 10 wild yeast per million yeast cells, (2) no more than 10 bacterial cell per million yeast cells, (3) no Salmonella in 25g product.

9. Toxicology

The petitioner provided chemical composition for three Merlot wines: (1) one produced using the ML01 yeast; (2) one produced using the parental yeast strain S92; and (3) one that underwent alcoholic fermentation using the strain S92 and malo-lactic fermentation using lactic acid bacteria (Oenococcus oeni). With the exception of ethyl lactate, the concentrations of other components were of similar magnitude for all 3 batches of wine. Due to the metabolic activity of ML01, higher levels of ethyl lactate in wines produced using ML01 are not unexpected since the bacterial malolactic fermentation was incomplete. No unique chemical products were detected in the ML01 derived wine.

The petitioner has indicated that, as a result of the clarification process, the wine will contain only the smaller polypeptides and amino acids from the two introduced gene sequences and that no intact sequences will remain. Wine that has undergone bacterial malolactic fermentation will also contain protein products of the gene that encodes the malolactic enzyme but not the protein products of the gene encoding malate permease. With respect to intact malate permease, the petitioner expects that it would not be present at the end of the alcoholic fermentation for two reasons: (1) suspected quick turnover of the permease; and (2) the observed absence of malate degradation after the end of the yeast's exponential growth phase.

Conclusion:

Health Canada's review of the information presented in support of the food use of wine yeast, Saccharomyces cerevisiae ML01 does not raise concerns related to food safety. Health Canada is of the opinion that wine derived from wine yeast ML01 is as safe as wine from current commercial wine yeast strains.

This opinion is solely with respect to the suitability of wine yeast ML01 for use in wine production. It is the continuing responsibility of Dr. van Vuuren and wine producers, to ensure that their products are in compliance with all applicable statutory and regulatory requirements. Any new information obtained in relation to these products which have potential health and safety implications should be forwarded to Health Canada for our consideration in order to ensure the continued safety and integrity of all foods available in the Canadian marketplace. The sale of a food which poses a hazard to the health of consumers would contravene the provisions of the Food and Drugs Act.

This Novel Food Information document has been prepared to summarize the opinion regarding the subject product provided by the Food Directorate, Health Products and Food Branch, Health Canada. This opinion is based upon the comprehensive review of information submitted by the petitioner according to the Guidelines for the Safety Assessment of Novel Foods.

(Également disponible en français)

For further information, please contact:

Novel Foods Section
Food Directorate
Health Products and Food Branch
Health Canada
Tunney's Pasture
Ottawa, Ontario K1A 0L2
Telephone:(613) 941-5535
Facsimile: (613) 952-6400