METABOLIC ENGINEERING OF INDUSTRIAL YEAST STRAINS TO MINIMIZE THE  PRODUCTION OF ETHYL CARBAMATE IN GRAPE AND SAKE WINE    by    MATTHEW SOLOMON DAHABIEH  B.Sc. Hon., The University of British Columbia, 2006      A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF   THE REQUIREMENTS FOR THE DEGREE OF    MASTER OF SCIENCE    in    THE FACULTY OF GRADUTATE STUDIES  (Genetics)          THE UNIVERISTY OF BRITISH COLUMBIA  (Vancouver)    April 2008  ©Matthew Solomon Dahabieh, 2008  ii    ABSTRACT    During alcoholic fermentation Saccharomyces cerevisiae metabolizes L‐arginine to ornithine and  urea. S. cerevisiae can metabolize urea through the action of urea amidolyase, encoded by the DUR1,2  gene; however, DUR1,2 is subject to nitrogen catabolite repression (NCR) in the presence of high quality  nitrogen sources during fermentation. Being cytotoxic at high concentrations, urea is exported into wine  where it spontaneously reacts with ethanol, and forms the carcinogen ethyl carbamate (EC).    Urea degrading yeast strains were created by integrating a linear cassette containing the DUR1,2  gene under the control of the S. cerevisiae PGK1 promoter and terminator signals into the URA3 locus of  the Sake yeast strains K7 and K9. The ‘self‐cloned’ strains K7EC‐ and K9EC‐ produced Sake wine with 68%  less EC. The Sake  strains K7EC‐ and K9EC‐ did not efficiently  reduce EC  in Chardonnay wine due  to  the  evolutionary adaptation of said strains to the unique nutrients of rice mash; therefore, the functionality  of  engineered  yeasts must  be  tested  in  their  niche  environments  as  to  correctly  characterize  new  strains.    S. cerevisiae possesses an NCR controlled high affinity urea permease  (DUR3). Urea  importing  yeast strains were created by integrating a linear cassette containing the DUR3 gene under the control  of the PGK1 promoter and terminator signals into the TRP1 locus of the yeast strains K7 (Sake) and 522  (wine).  In  Chardonnay wine,  the  urea  importing  strains  K7D3  and  522D3  reduced  EC  by  7%  and  81%,  respectively; reduction by these strains was equal to reduction by the urea degrading strains K7EC‐ and  522EC‐.  In  Sake  wine,  the  urea  degrading  strains  K7EC‐  and  522EC‐  reduced  EC  by  87%  and  84%  respectively, while the urea importing strains K7D3 and 522D3 were significantly less capable of reducing  EC (15% and 12% respectively). In Chardonnay and Sake wine, engineered strains that constitutively co‐ expressed DUR1,2 and DUR3 did not reduce EC more effectively than strains  in which either gene was  expressed  solely. Uptake  of  14C‐urea under non‐inducing  conditions was  enhanced  in urea  importing  strains; parental strains failed to  incorporate any 14C‐urea thus confirming the functionality of the urea  permease derived from the integrated DUR3 cassette.  iii    TABLE OF CONTENTS  ABSTRACT ...................................................................................................................................................... ii  TABLE OF CONTENTS .................................................................................................................................... iii  LIST OF TABLES .............................................................................................................................................. x  LIST OF FIGURES .......................................................................................................................................... xii  LIST OF ABBREVIATIONS .............................................................................................................................. xv  ACKNOWLEDGEMENTS .............................................................................................................................. xix  1 INTRODUCTION ......................................................................................................................................... 1  1.1 Saccharomyces cerevisiae ............................................................................................................ 1  1.2 Yeast strains from nature vs. laboratory yeasts ........................................................................... 3  1.3 S. cerevisiae and industry: Wine yeasts ....................................................................................... 5  1.4 Winemaking and Sake brewing .................................................................................................... 6  1.5 Aspergillus oryzae and its role in Sake brewing ......................................................................... 10  1.6 Yeast nitrogen metabolism during alcoholic fermentation ....................................................... 11  1.7 Nitrogen metabolism and Nitrogen Catabolite Repression (NCR) in S. cerevisiae .................... 12  1.8 Urea and ethyl carbamate (EC) .................................................................................................. 16  1.9 The EC problem .......................................................................................................................... 17  1.9.1 The EC problem: History ............................................................................................... 17  1.9.2 The EC problem: Surveys of EC in alcoholic beverages ................................................ 18  1.9.3 The EC problem: Current methods of lowering EC ....................................................... 19  1.9.3.1 Agricultural methods ....................................................................................... 19  1.9.3.2 Additives (acid urease) .................................................................................... 20  1.9.3.3 Additives (DAP) ................................................................................................ 20  iv    1.9.3.4 Genetic engineering (urease expression) ........................................................ 20  1.9.3.5 Genetic engineering (CAR1) ............................................................................. 20  1.9.3.6 Genetic engineering (DUR1,2) ......................................................................... 21  1.9.4 Alternative methods for EC reduction .......................................................................... 22  1.10 Introduction to DUR3: Role in the cell (urea transport, polyamines, boron) .......................... 23  1.11 Proposed Research ................................................................................................................... 25  1.11.1 Significance of Research ............................................................................................. 25  1.11.2 Hypotheses ................................................................................................................. 25  1.11.2.1  The  metabolically  engineered  Sake  yeast  strains  K7EC‐  and  K9EC‐  should  reduce EC efficiently during Sake brewing trials. ........................................... 25  1.11.2.2 Constitutive co‐expression of DUR1,2 and DUR3 in metabolically engineered  yeasts should result in synergistic EC reduction. ............................................ 26  1.11.3 Main objectives ........................................................................................................... 26  2 MATERIALS AND METHODS ..................................................................................................................... 27  2.1 Strains, plasmids and genetic cassettes ..................................................................................... 27  2.2 Culture conditions ...................................................................................................................... 28  2.3 Genetic construction of the urea degrading yeasts K7EC‐ and K9EC‐ ........................................... 29  2.3.1 Co‐transformation of the DUR1,2 cassette and pUT332 .............................................. 29  2.3.2 Screening of transformants for the integrated DUR1,2 cassette ................................. 29  2.3.3 Genetic characterization ............................................................................................... 30  2.3.3.1 Southern blot analyses .................................................................................... 30  2.3.3.2 Sequence analysis ............................................................................................ 30  v    2.3.3.3 Analysis of DUR1,2 gene expression by qRT‐PCR ............................................ 31  2.3.3.4 Global gene expression analysis ...................................................................... 32  2.3.4 Phenotypic characterization ......................................................................................... 33  2.3.4.1 Analysis of fermentation rate in Chardonnay must ........................................ 33  2.3.4.2 Analysis of fermentation rate in Sake mash .................................................... 33  2.3.4.3 Analysis of glucose/fructose utilization and ethanol production .................... 34  2.3.5 Functionality analyses ................................................................................................... 35  2.3.5.1 Reduction of EC in Chardonnay wine .............................................................. 35  2.3.5.2 Reduction of EC in Sake wine .......................................................................... 35  2.3.5.3 Quantification of EC in wine by solid phase microextraction and GC/MS ...... 35  2.4 Genetic construction of the urea importing yeasts K7D3, K7EC‐D3, 522D3, and 522EC‐D3 ............... 36  2.4.1 Construction of the DUR3 linear cassette ..................................................................... 36  2.4.1.1 Construction of pHVX2D3 ................................................................................ 36  2.4.1.2 Construction of pHVXKD3 ................................................................................ 37  2.4.1.3 Construction of pUCTRP1 ................................................................................ 38  2.4.1.4 Construction of pUCMD .................................................................................. 39  2.4.2 Sequence analysis of the DUR3 cassette in pUCMD ..................................................... 40  2.4.3  Transformation  of  the  linear  DUR3  cassette  into  S.  cerevisiae  and  selection  of  transformants .............................................................................................................. 43  2.4.3.1 Confirmation of integration via colony PCR .................................................... 43  2.4.4 Genetic characterization ............................................................................................... 43  vi    2.4.4.1 Southern blot analyses .................................................................................... 43  2.4.4.2 Analysis of gene expression by northern blotting ........................................... 44  2.4.4.3 Analysis of DUR3 gene expression by qRT‐PCR ............................................... 44  2.4.4.4 Global gene expression analysis ...................................................................... 45  2.4.5 Analysis of urea uptake using 14C‐urea ......................................................................... 45  2.4.6 Phenotypic characterization ......................................................................................... 45  2.4.6.1 Analysis of fermentation rate in Chardonnay must ........................................ 45  2.4.6.2 Analysis of fermentation rate in Sake mash .................................................... 45  2.4.6.3 Analysis for ethanol content  ........................................................................... 45  2.4.7 Functionality of metabolically enhanced yeasts ........................................................... 46  2.4.7.1 Reduction of EC in Chardonnay wine .............................................................. 46  2.4.7.2 Reduction of EC in Sake wine .......................................................................... 46  2.5 Statistical analyses........................................................................................................... 46  3 RESULTS ................................................................................................................................................... 47  3.1 Constitutive expression of DUR1,2 in Sake yeast strains K7 and K9 .......................................... 47  3.1.1  Integration of  the  linear DUR1,2 cassette  into  the genomes of Sake yeast strains K7  and K9 .......................................................................................................................... 47  3.1.2 Genetic characterization of K7EC‐ and K9EC‐ ................................................................... 47  3.1.2.1 Correct  integration of  the DUR1,2  linear cassette  into  the genomes of K7EC‐  and K9EC‐ .......................................................................................................... 47  3.1.2.2  Sake  strains  K7EC‐  and  K9EC‐  do  not  contain  the  bla  and  Tn5ble  antibiotic  resistance markers .......................................................................................... 50  vii    3.1.2.3 Sequence of  the DUR1,2  cassette  integrated  into  the genomes of K7EC‐ and  K9EC‐ ................................................................................................................. 51  3.1.2.4 Confirmation of constitutive expression of DUR1,2  in K7EC‐ and K9EC‐ by qRT‐ PCR .................................................................................................................. 52  3.1.2.5 Effect of  the  integrated DUR1,2 cassette on  the  transcriptomes of K7EC‐ and  K9EC‐ ................................................................................................................. 53  3.1.3 Phenotypic characterization of K7EC‐ and K9EC‐ ............................................................. 54  3.1.3.1 Fermentation rate of K7EC‐ and K9EC‐ in Chardonnay must .............................. 54  3.1.3.2 Fermentation rate of K7EC‐ and K9EC‐ in Sake mash .......................................... 55  3.1.3.3 Utilization  of  glucose  and  fructose  and  production  of  ethanol  by  K7EC‐  and  K9EC‐ in Sake wine ............................................................................................ 56  3.1.4 Constitutive expression of DUR1,2  in Sake yeast strains K7EC‐ and K9EC‐ reduces EC  in  Chardonnay wine by approximately 30% .................................................................... 57  3.1.5 Constitutive expression of DUR1,2  in Sake yeast strains K7EC‐ and K9EC‐ reduces EC  in  Sake wine by approximately 68% ................................................................................ 58  3.2 Constitutive expression of DUR3 in the Sake yeast strain K7 and the wine yeast strain 522 .... 59  3.2.1 Sequence of pUCMD ..................................................................................................... 59  3.2.2 Integration of the linear DUR3 cassette into the genomes of yeast strains K7, K7EC‐, K9,  and K9EC‐ ....................................................................................................................... 63  3.2.3 Genetic characterization of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 ....................................... 63  3.2.3.1 Correct  integration of  the DUR3  cassette  into  the genomes of K7D3, K7EC‐D3,  522D3, and 522EC‐D3 .......................................................................................... 63  3.2.3.2  Confirmation  of  constitutive  expression  of  DUR3  in  K7D3  and  K7EC‐D3  by  northern blotting ............................................................................................ 66  viii    3.2.3.3 Quantification of constitutive DUR3 expression in K7D3 and K7EC‐D3 by qRT‐PCR  ........................................................................................................................ 67  3.2.3.4 Effect of the integrated DUR3 cassette on the transcriptome of K7D3 ............ 68  3.2.4 Recombinant  strains K7D3 and K7EC‐D3 exhibit highly enhanced urea uptake ability  in  conditions of strong NCR ............................................................................................. 69  3.2.5 Recombinant  strains K7D3 and K7EC‐D3 exhibit highly enhanced urea uptake ability  in  conditions of NCR de‐repression ................................................................................. 72  3.2.6 Phenotypic characterization of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 ................................. 75  3.2.6.1 Fermentation rate of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 in Chardonnay wine ... 75  3.2.6.2 Ethanol production by K7D3, K7EC‐D3, 522D3, and 522EC‐D3 in Chardonnay wine . 77  3.2.6.3 Fermentation rate of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 in Sake wine ............... 77  3.2.6.4 Ethanol production by K7D3, K7EC‐D3, 522D3, and 522EC‐D3 in Sake wine ............. 79  3.2.7  Constitutive  expression  of  DUR3  in  yeast  strains  K7D3  and  522D3  reduces  EC  in  Chardonnay wine by 24.97% and 81.38%, respectively .............................................. 79  3.2.8 Constitutive expression of DUR3  in yeast  strains K7D3 and 522D3  reduces EC  in Sake  wine by 18.40% and 10.45%, respectively .................................................................. 80  4 DISCUSSION ............................................................................................................................................. 82  4.1 Constitutive expression of the DUR1,2 cassette reduces EC production in wine and Sake ...... 82  4.2 Integration of the DUR1,2 cassette  into the genomes of Sake yeast strains K7 and K9 yielded  the functional urea degrading Sake yeasts K7EC‐ and K9EC‐ ......................................................... 83  4.2.1 Genetic characterization of K7EC‐ and K9EC‐ ................................................................... 83  4.2.2 The Sake yeasts K7EC‐ and K9EC‐ conduct efficient alcoholic fermentations .................. 85  4.2.3 The Sake yeasts K7EC‐ and K9EC‐ reduce EC poorly in Chardonnay wine, yet efficiently in  Sake wine ..................................................................................................................... 85  ix    4.3  Constitutive  expression  of  the  urea  permease,  DUR3,  in  yeast  cells  is  a  viable  alternative  method to reduce EC in fermented alcoholic beverages ........................................................... 89  4.3.1 Construction of a  linear PGK1p‐DUR3‐PGK1t‐kanMX cassette  for  integration  into  the  TRP1 locus of wine and Sake yeasts ............................................................................ 89  4.3.2  Integration  of  the  DUR3  cassette  into  the  genomes  of  K7,  K7EC‐  ,  522,  and  522EC‐  yielded the functional urea transporting yeasts K7D3, K7EC‐D3, 522D3,  and 522EC‐D3 ..... 91  4.3.2.1 Integration of the DUR3 cassette into the genomes of K7D3, K7EC‐D3, 522D3, and  522EC‐D3 results in constitutive expression of DUR3 ........................................ 92  4.3.2.2 The integrated DUR3 cassette results in enhanced urea uptake .................... 93  4.3.2.3 The metabolically engineered yeasts K7D3, K7EC‐D3, 522D3, and 522EC‐D3 ferment  at similar  rates and produce similar amounts of ethanol  in Chardonnay and  Sake wine ........................................................................................................ 94  4.3.2.4  Variability  of  metabolically  engineered  yeasts  to  effectively  reduce  EC  in  Chardonnay wine and in Sake wine ................................................................ 95  4.3.2.5  The metabolically  engineered  yeasts  K7EC‐D3  and  522EC‐D3  do  not  reduce  EC  more effectively than K7EC‐ and 522EC‐ or K7D3 and 522D3 in either Chardonnay  or Sake wine .................................................................................................... 98  5 CONCLUSIONS ....................................................................................................................................... 101  5.1 Future Directions ...................................................................................................................... 103  REFERENCES ............................................................................................................................................. 104    x    LIST OF TABLES  Table 1.   Ranking of various yeast nitrogen sources according to NCR repression strength ............... 13  Table 2.   Maximum potential ethyl carbamate detected by GC/MS in 20 wines from six countries .. 19  Table 3.   Strains  used  in  the  genetic  construction  and  characterization  of  DUR3  expressing  yeast  strains. .................................................................................................................................... 27  Table 4.   Plasmids used  in  the genetic  construction and  characterization of DUR3 expressing yeast  strains ..................................................................................................................................... 28  Table 5.   Genetic cassettes used in the genetic construction and characterization of DUR3 expressing  yeast strains ........................................................................................................................... 28  Table 6.   Oligonucleotide primers used in sequencing of the integrated DUR1,2 cassette ................. 31  Table 7.   Oligonucleotide primers used in sequencing of DUR3 cassette in pUCMD .......................... 41  Table 8.   Discrepancies between the integrated DUR1,2 cassette of K9EC‐ and published sequences  51  Table 9.   Detailed description of the DNA sequences that comprise the DUR1,2 cassettes ............... 51  Table 10  Effect of  the  integrated DUR1,2  cassette  in  the genome of K7 on global gene expression  patterns in S. cerevisiae K7EC‐ (≥ 4‐fold change) ..................................................................... 53  Table 11.   Utilization of glucose and fructose and production of ethanol in Sake wine by parental yeast  strains (K7 and K9), their metabolically engineered counterparts (K7EC‐ and K9EC‐) .............. 57  Table 12.   Reduction of EC by functionally enhanced yeast strains K7EC‐ and K9EC‐ during wine making ...  ............................................................................................................................................... 58  Table 13.   Reduction of EC by functionally enhanced yeast K7EC‐ and K9EC‐ strains during Sake brewing ..  ............................................................................................................................................... 59  Table 14.   Discrepancies between the DUR3 cassette in pUCMD and published sequences ................ 60  Table 15.   Detailed description of the DNA sequences that comprise the DUR3 cassette .................... 60  xi    Table 16.   Recombinant yeast strains created by integration of the DUR3 cassette into the TRP1 locus  ............................................................................................................................................... 63  Table 17.  Effect  of  the  integrated  DUR3  cassette  in  the  genome  of  K7  on  global  gene  expression  patterns in S. cerevisiae K7D3 (≥ 4‐fold change) ..................................................................... 69  Table 18.  Ethanol produced by Sake yeast strains (K7, K7EC‐, K7D3, and K7EC‐D3) and wine yeast strains  (522, 522EC‐, 522D3, and 522EC‐D3) in Chardonnay wine ........................................................... 77  Table 19.   Ethanol produced by Sake yeast strains (K7, K7EC‐, K7D3, and K7EC‐D3) and wine yeast strains  (522, 522EC‐, 522D3, and 522EC‐D3) in Sake wine....................................................................... 79  Table 20.   Reduction of EC by functionally enhanced yeast strains during wine making ...................... 80  Table 21.   Reduction of EC reduction by functionally enhanced yeast strains during Sake making ...... 81    xii    LIST OF FIGURES  Figure 1.   Chemical basis of anaerobic fermentation .............................................................................. 5  Figure 2.   Simplified diagram of the Embden‐Meyerhof pathway .......................................................... 6  Figure 3.   Processes involved in red and white winemaking ................................................................... 8  Figure 4.   Contrasting processes in wine and Sake making ................................................................... 10  Figure 5.   Overview of urea metabolism in S. cerevisiae ....................................................................... 11  Figure 6.   Model of reciprocal regulation of GATA factor gene expression and GATA factor regulation  of NCR‐sensitive gene expression .......................................................................................... 14  Figure 7.   Permeases  and  degradative  enzymes  needed  to  utilize  poor  nitrogen  sources  are  transcriptionally silenced during growth in abundant high quality nitrogen sources ........... 14  Figure 8.   Model of the regulatory pathway by which rapamycin and nitrogen starvation  induce NCR  regulated gene expression ..................................................................................................... 16  Figure 9.   Synthesis reaction and bioactivation pathway of ethyl carbamate ...................................... 17  Figure 10.   Schematic representation of cloning strategy for creation of pHVX2D3 .............................. 37  Figure 11.   Schematic representation of cloning strategy for creation of pHVXKD3 .............................. 38  Figure 12.   Schematic representation of cloning strategy for creation of pUCTRP1 ............................... 38  Figure 13.   Schematic representation of cloning strategy for creation of pUCMD ................................. 40  Figure 14.   Schematic representation of pUCMD .................................................................................... 42  Figure 15.   Schematic representation of the linear DUR1,2 cassette ...................................................... 47  Figure 16.   Integration of the DUR1,2 cassette into the URA3 locus of K7EC‐ and K9EC‐ was confirmed by  Southern blot analyses using a DUR1,2 probe ....................................................................... 48    xiii    Figure 17.   Disruption of the URA3  locus by  integration of the DUR1,2 cassette  in K7EC‐ and K9EC‐   was  confirmed by Southern blot analyses using a URA3 probe ................................................... 49  Figure 18.   The  genetically  engineered  strains  K7EC‐  and  K9EC‐  do  not  contain  the  bla  and  Tn5ble  antibiotic resistance markers ................................................................................................. 50  Figure 19.   A  schematic  representation  of  new  ORFs  of  more  than  100  codons  generated  during  construction of the DUR1,2 cassette ..................................................................................... 52  Figure 20.   Gene expression analysis (qRT‐PCR) of K7, K7EC‐, K9, and K9EC‐  indicates functionality of the  DUR1,2 cassette and constitutive expression of DUR1,2 in non‐inducing (NCR) conditions .....  ............................................................................................................................................... 53  Figure 21.   Fermentation  profiles  (weight  loss)  of  parental  and  DUR1,2  engineered  Sake  strains  in  Chardonnay wine ................................................................................................................... 55  Figure 22.   Fermentation profiles (weight loss) of parental and DUR1,2 engineered Sake strains in Sake  wine........................................................................................................................................ 56  Figure 23.   DNA  sequence  alignment  of  S288C  and  pUCMD  revealed  nine  discrepancies  along  the  length of DUR3 cassette ........................................................................................................ 62  Figure 24.   A  schematic  representation  of  new  ORFs  of  more  than  100  codons  generated  during  construction of the DUR3 cassette ........................................................................................ 62  Figure 25.   Schematic representation of the linear DUR3 cassette ......................................................... 63  Figure 26.   Integration of the DUR3 cassette into the TRP1 locus of 522D3, 522EC‐D3, K7D3, and K7EC‐D3 was  confirmed by Southern blot analysis using DUR3 and TRP1 probes ..................................... 64  Figure 27.   Schematic  representation  of  the  signals  expected  during  Southern  blot  analysis  of  recombinant  yeasts  containing  the  recombinant DUR3  cassette  integrated  into  the  TRP1  locus ....................................................................................................................................... 65    xiv    Figure 28.   Alignment of the DNA sequences of S288C, 522, and K7 confirmed the presence of a mutant  EcoR1 site in the DUR3 coding region of K7. ......................................................................... 66  Figure 29.   Constitutive expression of DUR3 (2208 bp) was confirmed by northern blot analysis of K7,  K7EC‐, K7D3, K7EC‐D3 ................................................................................................................... 67  Figure 30.   Analyses of gene expression (qRT‐PCR) of K7, K7EC‐, K7D3, and K7EC‐D3 confirmed functionality  of  the  DUR3  cassette  and  constitutive  expression  of  DUR1,2  and  DUR3  in  non‐inducing  (NCR) conditions .................................................................................................................... 68  Figure 31.   Uptake of 14C‐urea by K7, K7EC‐, K7D3 and K7EC‐D3 under conditions of NCR ........................... 71  Figure 32.   Uptake of 14C‐urea by K7 and K7EC‐ under conditions of NCR ................................................ 72  Figure 33.  Uptake of 14C‐urea by K7, K7EC‐, K7D3 and K7EC‐D3 under conditions of NCR de‐repression .... 73  Figure 34.  Uptake of 14C‐urea by K7 and K7EC‐ under conditions of NCR de‐repression ......................... 74  Figure 35.   Fermentation profiles (weight  loss) of (a) Sake yeast strains K7, K7EC‐, K7D3, and K7EC‐D3 and  (b) wine yeast strains 522, 522EC‐, 522D3, and 522EC‐D3 in Chardonnay wine. ......................... 76  Figure 36.   Fermentation profiles (weight  loss) of (a) Sake yeast strains K7, K7EC‐, K7D3, and K7EC‐D3 and  (b) wine yeast strains 522, 522EC‐, 522D3, and 522EC‐D3 in Sake wine. ..................................... 78  Figure 37.   Schematic  representation  of  inducible  DUR1,2  expression  during  Chardonnay  and  Sake  wine fermentation by a urea importing yeast strain ............................................................. 97    xv    LIST OF ABBREVIATIONS  µg  Microgram  µM  Micromolar  aa  Amino acid  Abs  Absorbance  ANOVA  Analysis of variance  BC  British Columbia   bp  Base pair  cDNA  Complementary deoxyribonucleic acid  Ci  Curie  cm   Centimetre  cM  Centimorgan  Corp.   Corporation  cRNA  Ribonucleic acid derived from cDNA  DAP  Diammonium phosphate  DNA  Deoxyribonucleic acid  DTT  Dithiothreitol  EC  Ethyl carbamate  EDTA  Ethylenediamine tetraacetic acid  ER  Endoplasmic reticulum  EtOH  Ethanol  FDA  Food and Drug Administration  g  Gram  GC  Gas chromatograph  GC/MS  Gas chromatograph coupled mass spectrometry  GM  Genetically modified  GMO  Genetically modified organism  GO  Gene ontology  GRAS  Generally regarded as safe  HPLC  High pressure liquid chromatography  kb  Kilo base pair  xvi    kg  Kilogram  L  Litre  LB  Luria‐Bertani medium  LC  Liquid chromatograph  log  Logarithm  LSD  Least significant difference  m  Metre  M  Molarity  m/v  Mass per volume  mCi  Millicurie  mg  Milligram  min  Minute  mL  Millilitre  MLF  Malolactic fermentation  mM  Millimolar  mRNA  Messenger RNA  MS  Mass spectrometry  NAPS  Nucleic Acid Protein Service Unit at The University of British Columbia  NCR  Nitrogen catabolite repression system of S. cerevisiae  ng  Nanogram  nM  Nanomolar  nt  Nucleotide  °C  Degree Celsius  OD  Optical density  O/N  Overnight  ORF  Open reading frame  PB  Protein body  PCR  Polymerase Chain Reaction  PDM  Prise de Mousse  pH  Potential of Hydrogen  ppb  Parts per billion  xvii    ppm  Parts per million  qPCR  Quantitative PCR  qRT‐PCR   Quantitative Reverse Transcriptase PCR  RFLP  Restriction fragment length polymorphism  RNA  Ribonucleic acid  ROX  Passive reference dye used in qPCR  rpm  Revolutions per minute  RQ  Relative quantification  rRNA  Ribosomal RNA  RSD  Relative standard deviation  RT‐PCR  Reverse Transcriptase PCR  s  Second  SAP  Shrimp Alkaline Phosphatase  STDEV  Standard deviation  SGD  Saccharomyces Genome Database (www.yeastgenome.org)  SLR  Signal log ratio (log base = 2)  TBE  Buffer consisting of Tris base, boric acid, EDTA, and water  TE  Buffer consisting of Tris base, EDTA, and water  tRNA  Transfer RNA  TRP  Tryptophan  UAS  Upstream activation sequence  UASNTR  Upstream activation sequence – nitrogen regulated  UIS  Upstream induction sequence  UK   United Kingdom   URA  Uracil  URS  Upstream repression sequence  USA   United States of America   v  Volt  v/v  Volume per volume  w/o  Without  YAN  Yeast assimilable nitrogen   xviii    YEG  Medium consisting of yeast extract and dextrose  YNB  Yeast Nitrogen Base  YPD  Medium consisting of yeast extract, peptone, and dextrose  xix    ACKNOWLEDGEMENTS    It  is with great pleasure that  I thank the many people who contributed to this research and to  my development as a scientist. The road that scientists walk  is fraught with self‐doubt and frustration;  however, the people mentioned herein have been pivotal in helping me to complete my journey.    I  would  like  to  sincerely  thank  Dr.  Hennie  J.J.  van  Vuuren, my  research  supervisor,  for  his  support, guidance,  invaluable  insight, financial assistance, and help  in writing this thesis. Working with  Dr. van Vuuren at the UBC Wine Research Centre has allowed me to nurture my passion  for scientific  research while developing an appreciation  for the partnership between hypothesis‐driven science and  business. For the opportunities presented to me and the people I have met through Dr. van Vuuren, I am  exceptionally grateful.    I would like to thank the members of my supervisory committee, Dr. Ivan Sadowski and Dr. John  Smit for their advice, criticism, and diverse perspectives on this project. Working with Drs. Sadowski and  Smit was an  invaluable experience and  I am ever thankful of their willingness  to accommodate a very  tight completion timeline.    I would  like  to  thank Dr.  John  I. Husnik, a  former PhD  student  in  the van Vuuren  lab,  for his  patience,  care,  and  guidance.  John  was  an  indispensable  aide  for  resolving  problems,  clarifying  procedures,  and  sounding  ideas during my early development  as  a  scientist. He has  continued  to be  generous with his time and assistance despite living and working halfway across the country, and for this  I am thankful.    I would  like  to  thank my past and present Wine Research Centre colleagues, especially Calvin  Adams, Dr. Zongli Luo, and Lina Madilao for their warm friendship and valuable scientific collaboration.  My special  thanks are  in order  for Lina who completed all of  the GC/MS analysis  in  this study. To  the  members of  the  Food, Nutrition and Health administration office, Donna Bradley, Tram Nguyen, and  Patrick Leung,  I offer my  thanks  for  their help  in ordering and  logistics.  I would also  like  to  thank Dr.  Hugh Brock and Monica of the Genetics Graduate program (GGP) for their respective assistance during  my  time  in  the program. Additionally,  I would  like  to  thank my  fellow GGP  student and  friend, Kevin  Eade. Kevin’s friendship and coffee break companionship was much appreciated.   xx    I am extremely appreciative of those who funded this research and/or a portion of my studies:  National  Sciences  and  Engineering  Research  Council  of  Canada,  First  Venture  Technology  Corp.  (Vancouver, B.C.) and the Canadian Vintners Association.    Finally, and most importantly, I thank my parents, Elizabeth H. and Joseph Dahabieh. They bore  me, raised me, supported me, taught me, and loved me. Without them I would not be the person I am  today. 1    1  INTRODUCTION    1.1  Saccharomyces cerevisiae    In the early 20th century intensive genetic research began on the budding yeast Saccharomyces  cerevisiae  and,  ever  since,  scientists  have  championed  yeast  as  the  “Escherichia  coli  of  eukaryotes”  (Miklos and Rubin 1996). S. cerevisiae combines  the ease of use and brute  force genetics of bacteria  with  the  sophistication and elegance of higher eukaryotes,  thus making  it an  ideal organism  to  study  processes fundamental to all eukaryotic cells. Such processes, many of which are conserved from yeast  to  humans,  include  complex  cell  cycle  control,  eukaryotic  meiotic  recombination,  mitochondrial  respiration, and cell fusion events (Griffiths, et al. 2005).    S. cerevisiae is a unicellular eukaryotic fungus which is ubiquitous to wineries worldwide (Perez‐ Ortin,  Garcia‐Martinez  and  Alberola  2002). While  it  is  assumed  that  its  natural  environment  is  the  winery itself, S. cerevisiae can also be found naturally in rotting grapes and fruits, and in addition, there  may  be  some  yet  undiscovered  natural  habitat  of  budding  yeast  (Perez‐Ortin,  Garcia‐Martinez  and  Alberola 2002). S. cerevisiae  feeds on  the sugars and nutrients of crushed and rotting  fruit and, when  conditions are optimal, reproduce approximately every 90 minutes (Herskowitz 1988). Approximately 10  microns  in diameter,  yeast  reproduce  asexually by mitotic budding; however,  they  can  also undergo  sexual reproduction when haploid cells (created by meiotic division) of opposite mating type fuse, thus  yielding a stable diploid cell  (Herskowitz 1988). The yeast mating  types  (MATa and MATalpha) can be  likened to the male and female genders.    Wild isolates of S. cerevisiae also possess the ability to switch mating types such that a haploid  population  of  one  mating  type  can  mate  with  itself  and  achieve  diploidy  (Herskowitz  1988).  The  mechanism of mating  type  switching  involves  the HO endonuclease  and has been well  characterized  (Nasmyth 1993). Yeasts with this ability are known as homothallic and, from a Darwinian point of view,  achieving diploidy  is desirable. Having  two copies of every gene makes an organism genetically more  stable,  being  able  to  tolerate  loss  of  function  “recessive”  mutations  more  readily  than  a  haploid  equivalent (Greig and Travisano 2003). However, the haploid state can confer certain advantages on the  yeast cell; a second copy of every gene buffers deleterious mutations, but it also buffers advantageous  2    mutations. Thus, a diploid organism may, under certain circumstances, ‘evolve’ or ‘adapt’ to changes in  the environment more slowly than its haploid counterpart (Greig and Travisano 2003). Nevertheless, as  of present, S. cerevisiae has evolved into an organism which seems to function and prosper in a life cycle  stuck between its haploid prokaryotic predecessors and its diploid eukaryotic successors.    Like most bacteria and other microorganisms, yeast replicate rapidly and can be easily cultivated  in  the  laboratory. They grown well  in  liquid  culture and on  solid media, and  can be manipulated  via  standard microbiological  techniques  (Ausubel,  et  al. 2005).  In  addition  to other  common  techniques,  yeast  cells  are  remarkably  amenable  to  chemical or UV mutagenesis,  genetic  selection,  recombinant  DNA methods,  rescue  cloning,  complementation,  and  high  efficiency  transformation  (Griffiths,  et  al.  2005).  However,  the  real  value  of  S.  cerevisiae  lies  in  the  fact  that  yeasts  incorporate  all  of  these  advantages into a eukaryotic background. Thus, fundamental eukaryotic processes can be dissected on a  molecular level when such analysis would be exceedingly difficult or impossible in higher eukaryotes.    The full sequence of the S. cerevisiae genome was published in 1996 (Goffeau, et al. 1996) and  this has made S. cerevisiae even more powerful as a model organism. S. cerevisiae was  in fact the first  eukaryotic organism to be completely sequenced and subsequent analysis has revealed that the genome  of the common laboratory strain S288C is approximately 12 Mb in size and consists of 16 independently  assorting chromosomes (Goffeau, et al. 1996). It contains approximately 6000 genes, ~1000 of which are  essential  for  growth  on  rich media  (Maftahi, Gaillardin  and Nicaud  1998),  and  ~25%  of which  have  human  homologues  (Griffiths,  et  al.  2005).  The  remaining  5000  non‐essential  genes  have  been  systemically  knocked  out  in  the  ‘Saccharomyces  Genome  Deletion  project’  resulting  in  a  knock‐out  collection (Maftahi, Gaillardin and Nicaud 1998) that allows for streamlined reverse genetic analysis. In  addition to the set of 5000 knockouts available, the remaining 1000 essential genes can be investigated  through  the  use  of  readily  available  temperature  sensitive  (TS)  conditional  alleles  (Dohmen  and  Varshavsky  2005).  Finally,  both  expression  and  tiling DNA microarrays  exist  for  various  yeast  strains  allowing complex global analysis of the yeast genome, transcriptome and proteome (Dunn, Levine and  Sherlock  2005; Hauser,  et  al.  2001;  Perez‐Ortin, Garcia‐Martinez  and Alberola  2002; Rossignol,  et  al.  2003; Shobayashi, et al. 2007; Wu, et al. 2006).     3    Of  the  6000  yeast  genes,  the  Saccharomyces  Genome  Database  (SGD  –  http://www.yeastgenome.org) lists known functions or GO (Gene ontology) annotations for about 5000  genes.  The  remaining  1000  yeast  genes  remain  uncharacterized  despite  70  years  of  yeast  genetics.  While some speculate that many of these 1000 genes (uncharacterized ORFs) may not actually code for  functional protein, the general consensus  is that they are  indeed functional  (Pena‐Castillo and Hughes  2007). Upon in silico analysis, many of the 1000 uncharacterized genes contain putative protein domains  which suggest some sort of metabolic  function  (Pena‐Castillo and Hughes 2007). Furthermore, a  large  proportion of the 1000 uncharacterized genes are homologues to genes found solely in other species of  fungi  (Pena‐Castillo  and  Hughes  2007).  Thus,  these  genes  may  be  important  in  aspects  of  fungal  metabolism/physiology,  and  thus  do  not  assay well  under  standard  laboratory  conditions.  In  fact,  a  growing  consensus  amongst  yeast  biologists  is  that  these  1000  dubious  genes  will  never  become  characterized  until more  focus  is  placed  on  S.  cerevisiae  outside  the  laboratory  (Pena‐Castillo  and  Hughes 2007). This requires doing away with  traditional  laboratory screens and  looking at culturing S.  cerevisiae  under  natural  conditions,  most  notably  fermentative  conditions.  Under  conditions  of  fermentation,  yeast  cells  are  subjected  to  profoundly  different  stresses  than  under  laboratory  conditions.  These  stresses  include,  but  are  not  limited  to,  osmotic  stress,  ethanol  stress,  nutrient  limitation, oxidative stress, and temperature stress. If some of the 1000 uncharacterized genes are vital  to  these  types of stress  responses,  their  functions will only be  revealed when yeast cells are cultured  under  conditions  that  create  such  stresses.  For  example,  a  recent  study  of  the  yeast  transcriptome  during  wine  fermentation  identified  a  previously  uncharacterized  fermentation  stress  response  containing  approximately 223  genes, many of which  are part of  the 1000  remaining uncharacterized  yeast genes (Marks, et al. 2008).    1.2  Yeast strains from nature vs. laboratory yeasts    The majority of laboratory strains are descendents of isolates from nature, such as the common  laboratory strain S288C, which is a descendent of a yeast strain isolated from a rotten fig in California in  1938  (Perez‐Ortin,  Garcia‐Martinez  and  Alberola  2002).  However,  today’s  laboratory  strains  are  significantly different, both genetically and physiologically,  from  their wild  type parents. Given  the 90  minute generation time of S. cerevisiae, it is not difficult to imagine that over 70 years of growth on rich  laboratory media, a wild strain could evolve such that it no longer requires much of the genetic diversity  4    and  robustness needed  to deal with constantly changing environmental conditions  (Dunn, Levine and  Sherlock 2005). Without environmental pressures  to maintain  robust pathways needed  for growth  in  nature  (e.g.  sporulation,  pseudohyphal  growth),  natural  isolates  could  quickly  become  homogenized  into  the  less  vigorous  laboratory  strains  we  see  today.  As  such,  many  laboratory  strains  exhibit  dramatically different transcriptional profiles from wild yeasts and also differ in their ability to conduct  robust  and  efficient  alcoholic  fermentations  of  high  sugar  grape  musts  (Hauser,  et  al.  2001).  Additionally,  in  order  to maintain  stable  haploid  populations,  the  HO  locus  of  various  S.  cerevisiae  laboratory  strains  has  been  purposely  disrupted  (Nasmyth  1993).  Other  important  differences  in  laboratory yeast strains include the addition of various auxotrophic markers which can be used to select  transformants.  These  auxotrophic  markers  often  map  to  defects  in  amino  acid  or  nucleotide  biosynthetic pathways; common markers  include URA3, TRP1, LEU2, ADE2, etc.  (Ausubel, et al. 1995;  Dohmen and Varshavsky 2005).    In nature, many distinct strains of S. cerevisiae have been  isolated  from various environments  worldwide (Dunn, Levine and Sherlock 2005). Environments can vary widely in terms of nutrient (carbon,  nitrogen,  and minerals)  availability,  temperature,  osmolarity,  etc.  Thus, while  fundamentally  similar,  each of these strains has adapted, but not yet undergone speciation, in response to various niches in the  environment. These adaptations manifest themselves as differences  in growth rate, fermentation rate,  ethanol production, and various resistances when different strains are grown in identical media (Dunn,  Levine and Sherlock 2005; Hauser, et al. 2001).     Despite differences between wild  type strains  from nature,  they  tend  to share some common  distinguishing genetic characteristics when comparing them to  laboratory strains. As stated previously,  most  laboratory  strains  exist  as  stable  populations  of  haploids;  however,  most  wild  strains  are  homothallic and diploid, polyploid, or aneuploid  (Bond, et al. 2004; Dunn,  Levine and Sherlock 2005;  Hauser, et al. 2001; Hughes, et al. 2000; Perez‐Ortin, Garcia‐Martinez and Alberola 2002). Compared to  laboratory  strains,  wild  type  strains  also  differ  in  chromosome  length,  contain  large  scale  (~50  kb)  deletions or insertions, contain many more transposons (Ty elements), differ in sporulation rate (0‐75%)  and spore viability (0‐98%), exhibit variable pseudohyphal growth, and are largely heterozygous (Perez‐ Ortin, Garcia‐Martinez and Alberola 2002).  5    1.3  S. cerevisiae and industry: Wine yeasts    The  interaction between Homo sapiens and S. cerevisiae dates back almost 8000 years (Perez‐ Ortin, Garcia‐Martinez  and Alberola  2002;  Vine, Harkness  and  Linton  2002).  It was  first  reported  by  ancient Egyptian civilization that crushed grapes would ‘ferment spontaneously’ and that the resultant  wine contained  ‘magical and anesthetic properties’  (Vine, Harkness and Linton 2002). Since  that  time,  humans have been  selectively breeding  the  then unknown microorganism, S. cerevisiae,  for desirable  characteristics such as tolerance to high sugar stress, robust fermentation, ethanol tolerance, and good  flavour production (Hauser, et al. 2001).     As an art  form, winemaking  flourished  in  the ancient Mediterranean and was quickly adopted  anywhere where the climate was suitable for viticulture (Goode 2005; Vine, Harkness and Linton 2002).  As a science, however, winemaking was not understood until 1863; Louis Pasteur was the first to isolate  S. cerevisiae and show that  it was responsible for the production of ethyl alcohol (ethanol) and carbon  dioxide from simple sugars (glucose) (Figures 1 and 2) (Perez‐Ortin, Garcia‐Martinez and Alberola 2002;  Vine, Harkness and Linton 2002).      Sugar (glucose or fructose) → Ethanol + Carbon dioxide + ATP C6H12O6 + 2Pi + 2ADP → 2CH3CH2OH + 2CO2 + 2 ATP (energy released:118 kJ/mol) Figure 1. Chemical basis of  anaerobic  fermentation. Under  anaerobic  conditions  S.  cerevisiae  creates  energy (ATP) for biomass by converting sugar into ethanol and carbon dioxide.      6    Glucose Glycolysis 2 ADP 2 ATP 2 NAD+ 2 NADH 2 Ethanol 2 Acetylaldehyde 2 Pyruvate 2 CO2CH3CH2OH CH3(CO)H CH3(CO)COOH   Today, no other microorganism  is  as  important  to human diet  and  the  global economy  as  S.  cerevisiae (Goode 2005; Vine, Harkness and Linton 2002). Fermentation by yeast is vital to winemaking,  brewing, baking, and the distillation of spirits. Furthermore, in the face of global warming, S. cerevisiae  may soon play a vital role in the paradigm shift from fossil fuel dependency to the use of bio‐ethanol as  a fuel source (Farrell, et al. 2006).    1.4  Winemaking and Sake brewing    Although  it may be a  seemingly  simple process, winemaking  (enology) and  thus wine  itself  is  incredibly complex. Recent estimates suggest that wine contains approximately 1000 volatile flavor and  aroma compounds (Goode 2005). Furthermore, of the 1000 compounds, an estimated 400 are produced  by S. cerevisiae itself (Goode 2005). In its most fundamental form winemaking can be described as, “the  product  of  fermenting  [with  S.  cerevisiae]  and  processing  grape  juice  or must”  (Vine,  Harkness  and  Linton 2002).   Figure  2.  Simplified  diagram  of  the  Embden‐Meyerhof  pathway.  In  the  presence  of  oxygen,  pyruvate enters  the mitochondrial Krebs cycle and  then undergoes oxidative phosphorylation  to  create  maximal  ATP.  Under  anaerobic  conditions,  yeast  shunt  pyruvate  through  alcoholic  fermentation in order to create energy from sugar and restore the intracellular pool of NAD+ that is  depleted through glycolysis.    7      While human beings have been  actively  fermenting  grapes  and other  fruits  for  thousands of  years, the fundamental process has changed very little. The steps involved in modern grape winemaking  are outlined in Figure 3 (Vine, Harkness and Linton 2002).  8    Harvest of ripe grapes  (manual or machine)  De‐stemming and crushing of berries Treatment with pectolytic enzymes to   aid in clarification and juice extraction White or  Red?  Pressing of must to remove   skins and stems  Inoculation of must with commercial starter  S. cerevisiae culture (~150 available strains) or  allow natural flora yeasts to ferment  (Klockera, Hanseniaspora, Candida,   Metschnikowia, Pichia) Fermentation at 12‐18ºC   without mixing  Fermentation at 18‐30ºC  with mixing  MLF?  Malolactic fermentation (MLF) for deacidification of malate to lactate.  Requires addition of lactic acid bacteria (Oenococcus oeni)   or functionally enhanced S. cerevisiae strain ML01 capable of MLF   during alcoholic fermentation (Husnik, et al. 2006)   Anti‐microbial treatment  with sulfur dioxide (SO2)  Clarification by racking, fining,  centrifuging, and/or filtering  Oak? Maturation in stainless steel (most white wines)  Maturation in oak barrels   (most red wines and some  Chardonnays)  Blending (if desired) and bottling  White Red  No  Yes  YesNo                                                              Figure 3. Processes involved in red and white winemaking.  9    In contrast to grape wine, Sake wine is produced from the fermentation of milled rice grains by  S. cerevisiae. Sake is native to Japan but has steadily been gaining popularity around the world. Sake is  characterized by a translucent hue, mild fruity bouquet, relatively high natural alcohol content (15‐20%  v/v), and  is served either hot or cold  (Shobayashi, et al. 2007). As  is  the case with wine yeasts, many  different  Sake  yeast  strains  exist  and  certain  strains  have  become  popular  for  particular  sensory  characteristics  that  they  impart  to  the  final product. Some  strains produce Sake wine which  is  rich  in  fruitiness  and  has  high  acidity,  while  others  produce  wine  which  is  milder  and  more  aromatic  in  bouquet. Sake strains K7 and K9 are amongst the most popular and widely used yeasts  in the  industry  (Kodama  1993;  Wu,  et  al.  2006).  The  process  of  alcoholic  fermentation  is  fundamental  to  both  winemaking and Sake brewing and yeast cells are subjected  to substantial  temperature, acid, hypoxic  and ethanol stresses during both wine making and Sake brewing (Kodama 1993; Rossignol, et al. 2003;  Wu,  et  al.  2006).  However,  one  of  the  most  profound  differences  is  the  level  of  osmotic  stress  experienced by yeast cells during wine making. High ethanol levels produced during Sake fermentations  also exert significant ethanol stress on yeast cells.     Typical grape juice is a complex mixture high in carbohydrates, rich in assimilable nitrogen, and  high  in  vitamin/mineral  content  (Ingledew, Magnus  and  Patterson  1987;  Rossignol,  et  al.  2003). On  average,  grape must  contains  approximately  20%  w/v  sugar  (200  g/L)  in  the  form  of  a mixture  of  sucrose,  glucose  and  fructose  (Rossignol,  et  al.  2003).  Consequently,  at  the  start  of  grape  must  fermentation  yeast  cells  are  subject  to  substantial osmotic  stress. Although  yeast possess  a  cell wall  composed  of  crosslinked  1,3‐  and  1,6‐glucans  that  protects  them  against  osmotic  forces,  growth  in  grape must  induces  several other metabolic methods of  coping with  said  stress  (Westfall, Ballon and  Thorner 2004). The primary method of dealing with osmotic stress  is  induction of  the high osmolarity  growth (HOG) pathway (Reviewed in Westfall, Ballon and Thorner 2004; Han, et al. 1994). This pathway,  which  is  highly  conserved  between  most  eukaryotic  cells,  is  responsible  for  the  production  of  intracellular small molecules which help offset the osmotic pressure difference. The principle molecules  produced  in  S.  cerevisiae  are  glycerol  (1,2,3‐propanetriol)  and  trehalose  (α1,1  glucose  disaccharide)  (Westfall, Ballon and Thorner 2004). Both molecules are produced after  the  induction of biosynthetic  genes  by  the MAP  kinase mediated  HOG  pathway,  which  occurs  within minutes  of  osmotic  stress  (Westfall, Ballon and Thorner 2004). In contrast to the high osmolarity of typical grape must, Sake rice  10    mash contains much  less  initial free sugar as the majority of carbon  is tied up  in the form of  insoluble  starch (Kodama 1993; Shobayashi, et al. 2007; Wu, et al. 2006).     1.5  Aspergillus oryzae and its role in Sake brewing    Since  S.  cerevisiae  does  not  possess  the  necessary  amylases  needed  to  convert  starch  (β1,4  glucose polymer) to glucose, yeast cells are not able to consume starch as a sole carbon source (Kodama  1993;  Shobayashi,  et  al.  2007; Wu,  et  al.  2006).  As  a  result,  steamed  rice must  be  pre‐treated  in  preparation  for alcoholic  fermentation  (Figure 4). Sake brewers have  long used  the  fungus Aspergillus  oryzae as a source of amylases (Kodama 1993; Shobayashi, et al. 2007; Wu, et al. 2006).     At  the start of  fermentation steamed  rice  is mixed with  rice  that has been  inoculated with A.  oryzae, traditionally referred to as ‘koji’. Koji rice is rich in free glucose as well as amylases that are free  to act on the starch of freshly steamed rice. As a result of the presence of koji, glucose  is fed  into the  fermentation mixture  at  a  controlled  rate  (amylase  limited) which  substantially  lowers  the  osmotic  stress experienced by yeast cells (Kodama 1993; Shobayashi, et al. 2007; Wu, et al. 2006). This decrease  in osmotic stress allows more energy to be devoted to biomass and thus enables Sake fermentations to  reach higher titres and higher alcohol concentrations (Shobayashi, et al. 2007; Takagi, et al. 2005; Wu, et  al. 2006). Consequently, Sake wine  contains  the highest concentration of ethanol of any non‐distilled  alcoholic beverage (Kodama 1993).    Figure 4. Contrasting processes in wine and Sake making. During winemaking, a saccharification  step prior to alcoholic fermentation is unnecessary as the carbon source in grape must is mostly  glucose and fructose. During Sake brewing, saccharification must precede alcoholic fermentation  as S. cerevisiae cannot consume starch as a carbon source.    11    1.6  Yeast nitrogen metabolism during alcoholic fermentation    In order to build significant biomass and then conduct an efficient alcoholic fermentation, yeast  cells require significant amounts of nitrogen.   Free nitrogen,  in the  form of ammonia,  is used  in many  anabolic pathways, while peptides and free amino acids are either used in cellular processes directly, or  are broken down to ammonia, glutamate, and glutamine via catabolic pathways (Cooper 2002; Hofman‐ Bang 1999). The principle nitrogen source present  in grape must  is arginine, of which  the metabolism  leads directly to the formation of  intracellular urea (Figure 5) (Monteiro and Bisson 1991). Urea forms  from the arginase  (CAR1‐ EC 3.5.3.1) dependent breakdown of arginine to ornithine and urea  (Cooper  1982). At high concentrations, urea is a toxic and poor nitrogen source for S. cerevisiae, and is therefore  exported to the surrounding medium (Hofman‐Bang 1999). S. cerevisiae possesses the ability to degrade  urea via urea amidolyase (DUR1,2 ‐ EC 3.5.1.54) (Genbauffe and Cooper 1991); however, in the presence  of higher quality nitrogen sources, DUR1,2 expression is repressed while expression of the urea exporter  (DUR4)  is  not  (Whitney,  Cooper  and Magasanik  1973).  Consequently,  as  long  as  yeast  cells  are  not  starved  for  nitrogen, which  forces  them  to  degrade  urea,  they will  preferentially  export  urea  to  the  Figure 5 Overview of urea metabolism  in S. cerevisiae. Arginine  is  imported  into  the cell either by  the  Arginine  specific  transporter  CAN1  or  by  the  general  amino  acid  permease  GAP1.  Following  import  arginine is degraded to ornithine and urea by arginase, the product of the CAR1 gene. Urea can either be  exported by DUR4 or degraded to ammonia and carbon dioxide by DUR1,2. Cells can also reabsorb urea  through the importer DUR3. Adapted from Coulon, et al. (2006).    12    extracellular environment.  If,  in  the  later  stages of  fermentation,  yeast become  starved  for nitrogen,  urea  can  be  reabsorbed  (DUR3  ‐  TC  2.A.21.6)  (Cooper  and  Sumrada  1975;  ElBerry,  et  al.  1993)  and  metabolized; however, finished wines made from grape varietals with high assimilable nitrogen tend to  possess  significant  residual urea  (Ough, Crowell and Gutlove 1988; Ough, Crowell and Mooney 1988;  Ough, et al. 1990; Ough, et al. 1991).      1.7 Nitrogen metabolism and Nitrogen Catabolite Repression (NCR) in S. cerevisiae    The ability to discriminate between various nutrient sources  is an  important and evolutionarily  conserved theme  in biology. Prokaryotes and eukaryotes alike exhibit exquisite regulatory control over  their  metabolic  pathways  in  order  to  exist  in  the  most  energetically  efficient  way  possible.  Such  regulatory systems are often referred to as ‘catabolite repression’ systems because they repress genes  necessary to metabolize a particular nutrient source in the presence of a more favored one (Griffiths, et  al. 2005). Catabolite repression systems occur for both various carbon (Bruckner and Titgemeyer 2002;  Gancedo  1992)  and  nitrogen  sources  (Cooper  2002;  Hofman‐Bang  1999;  Salmon  and  Barre  1998).  Particularly well  characterized  examples  of  carbon  catabolite  repression  systems  include  the  lactose  (lac) operon  in E. coli (Griffiths, et al. 2005) as well as the galactose (GAL) genes  in S. cerevisiae (Lohr,  Venkov  and  Zlatanova  1995).  In  terms  of  nitrogen  catabolite  repression  both  prokaryotes  and  eukaryotes  utilize  a  more  global  repression  system  that  encompasses  many  different  catabolic  pathways.    Yeast nitrogen utilization  is centered on the usage of both glutamate and glutamine  (Hofman‐ Bang 1999). From glutamate and glutamine, wild type S. cerevisiae can synthesize any other amino acid  necessary  and  thus,  glutamate  and  glutamine,  along with  ammonia,  are  the  nitrogen  sources most  preferred by yeast (Hofman‐Bang 1999).    Due to  its diverse repertoire of nitrogen catabolic pathways, S. cerevisiae can grow solely on a  wide  range  of  nitrogenous  compounds  (e.g.  common  and  uncommon  amino  acids,  urea,  GABA,  allophanate, allantoin), as each of  these may be converted  in  to glutamate, glutamine, and ammonia  (Hofman‐Bang  1999). However,  because  each  non‐optimal  nitrogenous  compound  varies  in  terms  of  ease  of  import  and  degradation,  the  various  nitrogen  sources  can  be  ranked  in  terms  of  quality  or  13    preference (Hofman‐Bang 1999). Furthermore, since it is advantageous to utilize higher quality sources  preferentially, and because yeast posses a catabolite repression system for nitrogen sources, the various  compounds  can also be  ranked  in  terms of NCR  strength  (Hofman‐Bang 1999). A  few  common yeast  nitrogen sources are shown in Table 1.    Table 1. Ranking of various yeast nitrogen sources according to NCR repression strength. Low repression  strength indicates a poor nitrogen source. Adapted from Hofman‐Bang (1999).    NCR repression under  growth in listed media  (Low to High)  Nitrogen Source  1  Proline  2  GABA  3  Urea  4  Glutamate  5  Ammonium  6  Asparagine/Glutamine     The global nitrogen catabolite repression system of S. cerevisiae has been well studied but is far  from being understood  in absolute detail  (Reviewed  in Cooper 2002; Hofman‐Bang 1999). At  its core,  the NCR system of S. cerevisiae makes use of four known regulatory transcription factors (GLN3, GAT1,  DAL80,  and DEH1)  to  control  the  expression  of  all NCR  sensitive  genes  (Cooper  2002; Hofman‐Bang  1999).  Two  of  the  factors, GLN3  and GAT1,  are  positive  regulators  (activators), while  the  other  two  factors, DAL80 and DEH1, are negative regulators (repressors) (Cooper 2002; Hofman‐Bang 1999). The  complex  interaction of all four transcription factors, as well as various  inducers and repressors, at NCR  sensitive promoters allows for highly regulated nitrogen catabolite gene expression (Figures 6 and 7).    14        Figure 6. Model of reciprocal regulation of GATA factor gene expression and GATA factor regulation of  NCR‐sensitive gene expression. Dashed lines indicate weak association/regulation. Adapted from Cooper  (2002).      Figure  7.  Permeases  and  degradative  enzymes  needed  to  utilize  poor  nitrogen  sources  are  transcriptionally  silenced  during  growth  in  abundant  high  quality  nitrogen  sources.  Adapted  from  Cooper (2002).    15    The four NCR transcription factors are referred to as GATA factors because they all exert their  effect  though  the  DNA  consensus  sequence  5’‐GATAA‐3’.  Each  factor  binds  DNA  at  the  GATA  site  through a conserved zinc finger binding domain and each factor is homologous to many other zinc finger  proteins found  in higher eukaryotes,  including mammals (Bysani, Daugherty and Cooper 1991; Cooper  2002; Cox, et al. 2000; Cox, et al. 2004; Hofman‐Bang 1999; van Vuuren, et al. 1991).    The four NCR GATA factors are controlled by the upstream negative regulator URE2, which is in  turn regulated by the TOR pathway (Target of Rapamycin) (Cooper 2002; Hofman‐Bang 1999). The TOR  pathway, which  is  conserved  throughout most eukaryotic organisms, acts a master  regulatory  sensor  and signal transduction cascade that assesses and responds, with highly pleiotropic effects, to general  cell  health,  nutrient  availability,  and  growth  (Cooper  2002; Dann  and  Thomas  2006; De  Virgilio  and  Loewith 2006). The TOR pathway   acts primarily  through  two kinases, TOR1 and TOR2, and  regulates  fundamental cell functions such as the cell cycle, cell division, protein synthesis, etc. (Cooper 2002; Dann  and Thomas 2006; De Virgilio and Loewith 2006).    In terms of controlling NCR, the exact mechanism  linking the GATA transcription factors, URE2  and TOR1/2 has yet to be fully worked out. Despite confusion and conflicting data, researchers do agree  on the following relationships (Cooper 2002; Hofman‐Bang 1999):    a. GLN3p localization correlates highly with active transcription of NCR regulated genes  b. NCR regulated genes are largely activated when GLN3p is nuclear  c. URE2p complexes with GLN3p and the complex localizes to the cytoplasm  d. Inhibition of TOR1/2p correlates highly with decreased GLN3p phosphorylation,  which in turn  correlates with GLN3p being nuclear and NCR regulated genes being expressed    The most commonly accepted model of TOR1/2 control on NCR  is depicted  in Figure 8.  In this  model  TOR1/2  keep NCR  genes  repressed by  inhibiting  a phosphatase which  is necessary  for GLN3p  dephosphorylation and nuclear import.    16        Figure  8. Model  of  the  regulatory  pathway  by which  rapamycin  and  nitrogen  starvation  induce NCR  regulated gene expression. Adapted from Cooper (2002).       1.8  Urea and ethyl carbamate (EC)    The excretion of non‐metabolized urea into wine is the major factor involved in formation of the  compound ethyl carbamate (EC) (Monteiro and Bisson 1991).  Under ambient conditions (wine storage),  ethanol  reacts  with  carbamyl  compounds  present  in  fermented  beverages  to  form  EC  (Ingledew,  Magnus  and  Patterson  1987;  Kodama,  et  al.  1994; Ough,  Crowell  and  Gutlove  1988)  in  a  time  and  temperature dependent manner  (Figure 9). Potentially reactive carbamyl compounds  include citrulline  and carbamyl phosphate, which result from arginine and nucleotide metabolism, as well as urea (Ough,  Crowell and Gutlove 1988).   17        1.9  The EC problem    1.9.1  The EC problem: History    During the early part of the 20th century, ethyl carbamate was often administered to humans as  an anesthetic during surgeries. High  incidence of lung cancer in surgical patients was the first evidence  supporting EC’s  role  as  a  toxic  compound  (Nettleship, Henshaw  and Meyer 1943). Elucidation of  the  compound’s bioactivation pathway and further studies into its carcinogenicity (Ashby 1991; Guengerich  and Kim 1991; Leithauser, et al. 1990) fueled interest of both EC’s prevalence and mechanism of action.  Studies would soon show that vinyl carbamate epoxides, which are highly reactive oxidative degradation  products of EC,  interact with  free nucleotides as well as RNA and DNA and  induce mutations  in both  through mismatch pairing and direct damage (Dahl, Miller and Miller 1978; Leithauser, et al. 1990; Park,  et al. 1993; Schlatter and Lutz 1990; Zimmerli and Schlatter 1991). Further supporting EC’s role in cancer  is  the evidence  that EC  significantly  increases  the  rates of  liver,  lung, and harderian gland  cancers  in  male and female mice. Moreover,  incidence of mammary and ovarian as well as skin and forestomach  cancers was substantially increased in female and male mice, respectively (National Institutes of Health  National Toxicology Program 2004).     Figure 9. Synthesis reaction and bioactivation pathway of ethyl carbamate. a) The synthesis of EC  in wines results from the spontaneous reaction of ethanol and urea. b) The epoxide degradation  product  of  EC  binds  DNA,  causing  damage  and  resulting  in  increased  rates  of  cancers  in  test  animals.  Vinyl Carbamate Epoxide Vinyl CarbamateEthyl Carbamate a)  b)  18    Known  to  be  a  naturally  occurring  byproduct  of  fermentation,  EC  was  soon  shown  to  be  ubiquitous to nearly all wine and spirits albeit in vastly different quantities (Canas, et al. 1989; Ingledew,  Magnus and Patterson 1987; Ough 1976a; Ough 1976b; Schlatter and Lutz 1990; Zimmerli and Schlatter  1991). As a result of studies showing  that most wines and spirits contain high  levels of EC, during the  mid  1980’s  Canada  set  a  legal  limit  of  30  µg/L  on  the  allowable  EC  content  in wines;  the US  set  a  voluntary limit of 15 µg/L. Long term toxicology studies eventually gave rise to the 1997 action manual  on the prevention of EC in wine published by the FDA in conjunction with the Department of Viticulture  and  Enology  at  the University  of  California Davis  (http://vm.cfsan.fda.gov/~frf/ecaction.html)  (Butzke  and Bisson 1998).         Exposure  to EC, which may be  significantly  increased by  the  regular consumption of alcoholic  beverages  (Zimmerli  and  Schlatter  1991),  may  be  a  significant  factor  involved  in  human  cellular  mutagenesis and resultant tumorigenesis. As a result, winemakers have been actively reducing EC levels  in wines  both  by  agricultural  practices  and, more  recently, molecular  biological means  (Butzke  and   Bisson 1998).    1.9.2  The EC problem: Surveys of EC in alcoholic beverages    Despite  the  government  imposed  limits  on  EC  levels,  table wines,  Sake  and  other  alcoholic  beverages currently available to consumers contain far more EC than was previously suggested.     In order  to  assess  the  scope of  the EC problem, numerous  stuides have  assayed EC  levels  in  various  foods  and beverages. The  EC  content of 20  randomly  chosen wines  from  six wine producing  countries (five whites – Riesling, Pinot Gris, and Chenin Blanc; 15 reds – Pinot Noir, Cabernet Sauvignon,  Shiraz, Zinfandel, and Nebbiolo)  is summarized  in Table 2  (Reproduced  from Coulon, et al.  (2006)). As  the data indicates, of the 20 wines, 14 exceeded the Canadian EC legal limit (30 µg/L) and 17 exceeded  the US voluntary limit (15 µg/L) (Coulon, et al. 2006).     19      Table 2. Maximum potential ethyl carbamate detected by GC/MS in 20 wines from six countries. Wines  were heated at 70°C for 48 hours prior to analysis. Adapted from Coulon et al. (2006).  Wine Concentration EC (µg/L) Area 1 Area 2 Area 3 Mean peak area SD RSDa (%) 1 11.83 4274 4542 4185 4334 186 4.29 2 9.12 3649 3586 3887 3707 159 4.28 3 14.63 5043 4959 4948 4983 52 1.04 4 38.64 10262 10112 11269 10548 629 5.97 5 57.57 14342 15833 14632 14936 791 5.29 6 66.15 16610 16734 17430 16925 442 2.61 7 59.52 15890 15015 15260 15388 451 2.93 8 46.39 11208 13611 12213 12344 1207 9.78 9 55.41 13865 14820 14618 14434 503 3.49 10 24.36 7012 6808 7895 7238 578 7.98 11 44.76 11650 12576 11677 11968 527 4.40 12 41.32 11186 10952 11368 11169 209 1.87 13 27.5 7664 8366 7868 7966 361 4.53 14 38.22 10375 10471 10508 10451 69 0.66 15 57.68 14069 14839 15974 14961 958 6.41 16 33.82 9034 9428 9832 9431 399 4.23 17 51.99 14146 12880 13899 13642 671 4.92 18 18.61 5734 5935 6047 5905 159 2.69 19 34.99 9396 9295 10417 9703 621 6.40 20 36.57 9635 9756 10811 10067 647 6.43   a – Relative standard deviation    In  addition  to  the  goal of  reducing EC  in  table wines,  Sake wine has  some of  the highest EC  contents amongst fermented beverages, thus making the reduction of EC  in Sake an  intensely relevant  and important pursuit. Typical Sake wines have anywhere from 100‐250 µg/L EC (Canas, et al. 1989) due  to a pasteurization process that all Sakes undergo prior to bottling.    1.9.3  The EC problem: Current methods of lowering EC    Current strategies for EC reduction generally fall into three categories: agricultural practices, the  use of wine additives, and genetic engineering of yeasts.    1.9.3.1    Agricultural  methods.  The  yeast  assimilable  nitrogen  (YAN)  and, more  specifically,  arginine  content of the fermentation substrate directly influences the amount of EC present in the final product  (Butzke and Bisson 1998). Thus, a reasonable approach to controlling EC levels is to limit arginine levels  20    in grape must and rice mash. It  is possible to regulate the type and amount of nitrogen  in fertilizers  in  order  to minimize  urea  production.  Furthermore, management  of  legume  ground  cover  foliage  and  nitrogen fixing bacteria can keep juice arginine content below 1000 mg/L, however additional methods  are needed to further reduce EC levels in wines (Butzke and Bisson 1998).     1.9.3.2  Additives (acid urease). Alternative methodologies for EC reduction include the use of post‐ or  peri‐fermentation additives  (Ough and Trioli 1988). These additives, which are most often  lyophilized  preparations of urease from Lactobacillus fermentum, degrade urea  in the wine before  it has a chance  to form EC; however, urease additives yield variable results due to pH and ethanol sensitivity (Kodama,  et al. 1994) which, can significantly lengthen wine processing time due to necessary enzyme incubation.  The enzyme is also expensive for winemakers to use.    1.9.3.3    Additives  (DAP).  Supplementation  of  grape musts with  nitrogen  is  a  common winemaking  practice.  In musts with  low  assimilable  nitrogen,  supplementation  is  crucial  for  preventing  stuck  or  sluggish  fermentations  and  subsequent  spoilage.  Given  the  obvious  inappropriateness  of  supplementation with urea or arginine, the  industry makes use of the cheap and efficient supplement  diammonium phosphate  (DAP); DAP  is a  source of  free ammonia and, as a highly  favorable nitrogen  source,  has  been  shown  to  downregulate  CAR1  during  alcoholic  fermentation  (Marks,  et  al.  2003).  Lessening the cell’s dependence on arginine as a nitrogen source could prove to be a useful method for  lowering EC content in wines.    1.9.3.4  Genetic engineering (urease expression). Another possible method for EC control in wine is the  engineering  of  yeast  strains  which  express  bacterial  or  fungal  ureases  (Ough  and  Trioli  1988).  This  approach, whether  integrated  or  plasmid  borne, would  allow  yeast  to  directly  degrade  urea  during  fermentation,  and  eliminates  the  need  for  urease  additives  post‐fermentation. However,  because  S.  cerevisiae does not possess such an enzyme, a urease expressing strain would be classified as transgenic  thus  resulting  in  difficulty  obtaining  regulatory  approval  for  use.  Furthermore,  transgenic  organisms  carry a strong negative connotation in today’s society thus making their universal acceptance unlikely.    1.9.3.5    Genetic  engineering  (CAR1). More  central  to  this work  is  the  genetic  engineering  of  yeast  strains which are disrupted at the CAR1  locus (Kitamoto, et al. 1991). By disrupting CAR1, arginase will  no  longer be available to degrade arginine to urea thus preventing the formation of EC  in wine. While  21    this method does work  in  the  laboratory  (Kitamoto, et al. 1991; Yoshiuchi, Watanabe and Nishimura  2000), it has not been widely adopted in practice because arginine is one of the major nitrogen sources  for  S.  cerevisiae  (Ingledew, Magnus  and  Patterson  1987; Rossignol,  et  al.  2003). Moreover,  usage  of  Δcar1 strains is difficult due to the diploid nature of all industrial wine and Sake yeasts, and because of  the  high  risk  of  contamination  from  wild  type  CAR1/CAR1  yeasts;  however,  the  problem  of  contamination has been dealt with  in the case of Sake yeast by engineering Δcar1 with killer character  (Yoshiuchi, Watanabe  and Nishimura 2000).  It  should be noted  that CAR1  knockouts have only been  attempted in Sake yeast since it is generally regarded that yeast’s inability to metabolize arginine would,  in most cases, lead to stuck fermentations and possible spoilage. An antisense mediated method for the  knockdown of CAR1 has been successful in reducing Arginase activity in laboratory yeasts (Park, Shin and  Woo 2001); however, this methodology has yet to be examined in industrial yeasts.    1.9.3.6   Genetic  engineering  (DUR1,2).  Yeast  possess  the  ability  to  degrade  urea  via DUR1,2 which  encodes a bi‐functional adenosine tri‐phosphate (ATP) and biotin dependent enzyme (urea amidolyase)  that degrades urea to two molecules each of CO2 and NH3 in a two step reaction (Genbauffe and Cooper  1986; Genbauffe  and Cooper  1991; Whitney  and Cooper  1972; Whitney  and Cooper  1973; Whitney,  Cooper and Magasanik 1973). Urea is first carboxylated using ATP and biotin to form allophanate, after  which allophanate is hydrolyzed to form CO2 and NH3.     Since DUR1,2  is subject to regulation by NCR, and because regulatory mechanisms exist which  allow for production of wines with high residual urea content, it is reasonable to expect that constitutive  expression of DUR1,2 should result  in substantially  lowered EC  levels.  Indeed, our group has explored  this approach, and such yeasts are capable of producing wines which contain up to 89% less EC (Coulon,  et al. 2006). More specifically, when the DUR1,2 ORF was placed under the control of the yeast PGK1  promoter and terminator signals and a single copy was integrated into the URA3 locus of the industrial  strain UC Davis 522, Chardonnay wine created by the metabolically engineered strain (522EC‐) contained  89.1%  less EC  (Coulon, et al. 2006). Analysis of  the genotype, phenotype and  transcriptome of 522EC‐  suggested  that  the metabolically  engineered  strain  was  substantially  equivalent  to  its  parent,  thus  making  it  suitable  for commercialization  (Coulon, et al. 2006). Furthermore,  since  the urea degrading  strain  (522EC‐) contains no  foreign DNA sequences,  it  is not classified as  transgenic and  thus has been  22    given FDA GRAS approval which should make it much more readily accepted by industry and consumers  (Coulon, et al. 2006).    1.9.4  Alternative methods for EC reduction    Realistically,  there  are  only  four  genes  that  can  be manipulated  to  reduce  EC  in wine:  urea  production (CAR1), urea degradation (DUR1,2), urea export (DUR4), and urea import (DUR3).     Based on  the  literature and published data, we understand  the effects of manipulating CAR1  and DUR1,2 expression (Coulon, et al. 2006; Yoshiuchi, Watanabe and Nishimura 2000), and while CAR1  mutants produce almost no EC, CAR1 is not a practical industrial target for EC reduction due to problems  with stuck fermentations. Thus, manipulation of DUR4 and DUR3 remain as potential targets to reduce  EC  in wine and  Sake.   One option would be  to  knockout  the urea exporter DUR4,  thus disabling  the  yeast’s ability to export urea into the wine. Although this sounds like an attractive option, this strategy  would  likely cause serious problems for winemakers. Urea  is a toxic byproduct of arginine metabolism  and yeast cells need to export urea in order to stay healthy. By knocking out DUR4, urea will accumulate  in  the  cell  and  lead  to  an  adverse  physiological  state  resulting  in  stuck  fermentations.  The  fact  that  existing  recombinant DUR1,2  yeasts  produce  any  EC  at  all  (Coulon,  et  al.  2006)  suggests  that  under  typical winemaking  conditions more urea  is produced  than DUR1,2p,  from a  constitutively expressed  copy of DUR1,2, can degrade.     Given the potential problems associated with DUR4 knockouts, we focused our attention on the  urea permease, DUR3. Constitutive expression of DUR3 should result  in yeasts which will act  like urea  sponges;  not  only  will  these  yeasts  reabsorb  urea  that  they  excreted  as  a  byproduct  of  arginine  metabolism,  but  they  should  also  absorb  a  significant  amount  of  urea  that  is  naturally  present  in  fermentation substrate. By combining DUR1,2 constitutive expression with  that of DUR3,  it should be  possible to create recombinant yeasts which can conduct efficient and substantially equivalent alcoholic  fermentations with the production of little or no EC.        23    1.10  Introduction to DUR3: Role in the cell (urea transport, polyamines, boron)    During the early 1970’s it was known that yeast cells were capable of metabolizing urea as a sole  nitrogen source (Cooper and Sumrada 1975); however, there was no detailed knowledge of how it was  brought into the cell. Studies using 14C‐urea revealed that entry of urea into the cell is bimodal (Cooper  and Sumrada 1975; Sumrada, Gorski and Cooper 1976). A  facilitated diffusion system brings urea  into  the cell in an energy independent fashion when urea is present at concentrations greater than 0.5 mM.  More interestingly however, is the presence of an energy (ATP) dependent active transport system (Km =  14 μM) which functions at low concentrations of urea and is sensitive to nitrogen catabolite repression  (Cooper and Sumrada 1975; Sumrada, Gorski and Cooper 1976).    First purified and characterized in 1993, the DUR3 (Degradation of URea) ORF encodes a 735 aa  integral membrane protein which contains 15 predicted transmembrane domains (ElBerry, et al. 1993).  The protein, which localizes to the plasma membrane, utilizes ATP to transport urea into the cell at low  extracellular concentrations. There is some evidence to suggest that the physiological functioning state  of  the urea  transporter  is a multimeric complex, however  this has not been confirmed  (ElBerry, et al.  1993).    Expression  of  DUR3  is  regulated  in  a manner  highly  similar  to  other  genes  in  the  urea  and  allantoin degradative pathways (ElBerry, et al. 1993). Being subject to NCR, the DUR3 promoter contains  two  sets  of  tandem  GATAA  consensus  sequences;  however,  while  the  promoter  contains  efficient  GATAA  transcription  factor binding  sites  (UASNTR), high  level expression  is  strongly dependent on  two  upstream  induction  sequences  (UIS)  (Hofman‐Bang 1999). During growth on proline media  (no NCR),  little  DUR3  (and  DUR1,2)  mRNA  can  be  detected  by  northern  blotting  without  the  presence  of  a  gratuitous inducer (oxalurate, an allophanate analogue) (Hofman‐Bang 1999).    The activating transcription factors DAL81 and DAL82 act through the UIS sequences contained  in the DUR3 promoter. The consensus sequence, 5’ (G/C) AAA (A/T) NTGCG (T/C) T (T/G/C) (T/G/C) 3’,  for DAL81 and DAL82 binding  is shared between other allophanate  induced genes (CAR2, DAL2, DAL4,  DUR1,2 and DUR3 (Hofman‐Bang 1999).     24    During  times of  strong NCR, DUR3  is  actively  repressed by  the negative GATAA  transcription  factor DAL80, and deletion of DAL80 results  in expression of DUR3 even  in  the absence of an  inducer  (Hofman‐Bang 1999). Conversely, during NCR de‐repression, DUR3 is actively transcribed, if an inducer is  present, through the actions of the positive GATAA transcription factor GLN3 (Hofman‐Bang 1999).    In addition  to  the obvious  role  in urea uptake, DUR3 has been shown  to be  involved  in other  important cellular processes. DUR3 has been shown to be an important regulator of intracellular boron  concentration  (Nozawa,  et  al.  2006).  Cells  lacking  DUR3  show  decreased  intracellular  boron  concentration; thus, DUR3 appears to function as an active transporter of boron into the cell (Nozawa,  et al. 2006). Although evidence suggests DUR3 plays a  role  in boron  transport and  regulation, a clear  physiological role for DUR3 in terms of boron utilization has yet to be defined.     The  other  important  role  of  DUR3  is  in  the  uptake  of  polyamines  (Uemura,  Kashiwagi  and  Igarashi 2007). Polyamines, such as putrescine, spermidine, and spermine, are highly regulated peptides  essential  for  cell  growth  and  proliferation  (Uemura,  Kashiwagi  and  Igarashi  2007).  Their  function  is  ubiquitous  to  both  pro‐  and  eukaryotes.  Like  E.  coli,  S.  cerevisiae  possesses  general  polyamine  transporters  (TPO1‐4,  UGA4,  TPO5,  GAP1),  as  well  as  polyamine  specific  transporters  (AGP2).  Interestingly, DUR3 has been shown to specifically uptake polyamines concurrently with urea (Uemura,  Kashiwagi and  Igarashi 2007). As urea  is a very poor nitrogen source, and does not normally occur  in  significant  quantities  outside  the  cell,  the main  physiological  role  of  DUR3 may  well  be  polyamine  uptake;  in  fact, DUR3 mRNA  is  repressed  in  the presence of  large quantities of polyamines  (Uemura,  Kashiwagi and Igarashi 2007).    Most interesting is the apparent post‐translational regulation of DUR3 polyamine uptake by the  serine/threonine kinase PTK2 (Uemura, Kashiwagi and Igarashi 2007). PTK2 seems to positively regulate  DUR3 polyamine uptake via the phosphorylation of cytoplasmic residues Thr‐250, Ser‐251, and Thr‐684  (Uemura,  Kashiwagi  and  Igarashi  2007).  Although  DUR3  polyamine  activity  and  subsequent  PTK2  regulation has been preliminarily  investigated  in the  laboratory yeast YPH499 (Uemura, Kashiwagi and  Igarashi 2007), there are no known studies which have investigated the role of DUR3 mediated urea or  polyamine  uptake  during  alcoholic  fermentation. However,  it  is  known  that  different  strains  of wine  25    yeast react differentially  in terms of fermentation rate and biomass production  in response to varying  polyamine concentrations (Uemura, Kashiwagi and Igarashi 2007).     As with polyamine uptake, the DUR3 mediated uptake of boron seems to be post‐translationally  regulated. Although cells lacking DUR3 exhibit lower boron concentrations, the converse situation is not  true i.e. cells overexpressing DUR3 do not show significant increases in boron concentration (Nozawa, et  al.  2006).  Taken  together,  the  cases  of  polyamine  and  boron  uptake  provide  good  evidence  for  the  existence of a DUR3 regulatory protein, presumably PTK2.    1.11  Proposed Research  1.11.1  Significance of Research    Given  the obvious  governmental health  concern  regarding  the  EC  content of wines,  it  seems  logical  to  pursue  the  goal  of  producing wines  that  contain  no  EC. Until  recently with  the  advent  of  constitutive  DUR1,2  expression  (Coulon,  et  al.  2006),  existing  methods  of  EC  reduction  were  cumbersome, ineffective, expensive, and/or impractical.     The development of non‐transgenic yeast strains which are capable of producing little or no EC  during  an  efficient  and  substantially  equivalent  alcoholic  fermentation would  be  of  direct  benefit  to  industry and consumers alike. As such, the research described herein  is both an application of existing  technology to a novel target, as well as a proof of concept exploration of one possible new method for  EC reduction.    1.11.2  Hypotheses    1.11.2.1   The metabolically engineered Sake yeast strains K7EC‐ and K9EC‐ should reduce EC efficiently  during Sake brewing trials. Given the substantially different environment of Sake mash, it is reasonable  that the EC reduction of functionally enhanced Sake yeasts will be highly superior when fermenting rice  mash  rather  than  grape  must.  Sake  yeasts  have  evolved  to  function  optimally  in  the  nutrient  composition of rice mash and in the presence of ‘koji’, thus the efficiency of the DUR1,2 cassette should  also function optimally. Such a result should reveal the true EC reduction potential of our recombinant  26    yeasts  and  would  affirm  our  belief  that  each  specific  yeast  strain  must  be  tested  in  its  native  environment in order to yield the most accurate results.  1.11.2.2   Constitutive co‐expression of DUR1,2 and DUR3  in metabolically engineered yeasts should  result in synergistic EC reduction. DUR1,2/DUR3 yeasts should produce substantially  less EC than both  DUR1,2 and parental yeasts due  to  their ability  to absorb native urea  in grape musts and  to reabsorb  excreted  urea. Moreover,  these  yeasts  should  behave  like  their  parental  counterparts  in  all  other  aspects of  fermentation  i.e. growth  rate, ethanol production, CO2 production, kinetics, etc. Obtaining  these results will affirm our belief that the problem with current DUR1,2 clones is their ability to export  metabolic urea before  it  can be  completely degraded. Additionally DUR1,2  yeasts  cannot be used  to  degrade any urea natively present in the must/mash while DUR1,2/DUR3 yeasts could.    1.11.3  Main objectives    The main objectives of this study are to:    1. Constitutively express DUR1,2 in the Sake yeast strains K7 and K9, characterize the resultant  engineered  strains,  and  evaluate  the  effect  of  constitutive  DUR1,2  expression  on  EC  production in Chardonnay wine    2. Characterize  the  EC  reduction  potential  of  Sake  yeast  clones  K7EC‐  and  K9EC‐  during  small  scale Sake fermentation  3. Construct a genetic cassette capable of maintaining constitutive expression of a functional  DUR3 urea permease in wine and Sake yeasts  4. Constitutively express DUR3 both on its own and concurrently with DUR1,2 during wine and  Sake making in order to assess the effect on EC reduction  5. Characterize the role of DUR3, and its interplay with DUR1,2, in yeast urea metabolism and  EC production during alcoholic fermentation.      27    2  MATERIALS AND METHODS    2.1 Strains, plasmids and genetic cassettes      The  strains, plasmids,  and  genetic  cassettes used  in  the  construction  and  characterization of  DUR1,2 and/or DUR3 expressing yeast strains are listed in Tables 3, 4, and 5, respectively.    Table 3. Strains used in the genetic construction and characterization of DUR3 expressing yeast strains.  Strain  Description  Reference  E. coli Subcloning  Efficiency™ DH5α™  Competent Cells  F‐ φ80lacZΔM15 Δ(lacZYA‐argF)U169 recA1  endA1 hsdR17(rk‐, mk+) phoA  supE44 thi‐1 gyrA96 relA1 λ‐  Invitrogen  S. cerevisiae K7  Industrial Sake yeast strain Kyokai No. 701 (K7)  Brewing Society of Japan  S. cerevisiae K9  Industrial Sake yeast strain Kyokai No. 901 (K9)  Brewing Society of Japan  S. cerevisiae 522   Industrial wine yeast strain   UC Davis  S. cerevisiae K7EC‐  Sake yeast strain K7 containing the DUR1,2  cassette (Coulon, et al. 2006) integrated at the  URA3 locus  This study  S. cerevisiae K9EC‐  Sake yeast strain K9 containing the DUR1,2  cassette integrated at the URA3 locus  This study  S. cerevisiae 522EC‐  Wine yeast strain 522 containing the DUR1,2  cassette integrated at the URA3 locus  (Coulon, et al. 2006)  S. cerevisiae K7D3  Sake yeast strain K7 containing the DUR3  cassette integrated at the TRP1 locus  This study  S. cerevisiae 522D3  Wine yeast strain 522 containing the DUR3  cassette integrated at the TRP1 locus  This study  S. cerevisiae K7EC‐D3  Sake yeast strain K7 containing the DUR1,2  cassette integrated at the URA3 locus and the  DUR3 cassette integrated at the TRP1 locus  This study  S. cerevisiae 522EC‐D3  Wine yeast strain 522 containing the DUR1,2  cassette integrated at the URA3 locus and the  DUR3 cassette integrated at the TRP1 locus  This study  A. oryzae ‘Koji‐Kin’  Industrial preparation of ‘koji’ grade A. oryzae   Vision Brewing  http://www.visionbrewing.com              28    Table  4.  Plasmids  used  in  the  genetic  construction  and  characterization  of  DUR3  expressing  yeast  strains.  Plasmid  Description  Reference  pUT332∆ura3  E. coli/S. cerevisiae episomal shuttle vector containing the  Tn5ble (phleomycin) dominant marker.  (Gatignol 1987)  pUC18  High copy number E. coli plasmid that contains AmpR, ori,  and an MCS located within a lacZ coding sequence thus  facilitating cloning via blue/white X‐Gal selection.  (Yanisch‐Perron, Vieira  and Messing 1985)  pUG6  High copy number E. coli plasmid that contains AmpR,  kanMXR, and ori.  (Guldener, et al. 1996)  pHVX2  A YEplac181 based S. cerevisiae expression vector in which  gene expression is driven from the constitutive PGK1  promoter and terminator signals  (Volschenk, et al. 1997)  pUCTRP1  pUC18 to which the TRP1 coding region was inserted into  the BamH1 site at the MCS  This study  pHVX2D3  pHVX2 to which the DUR3 coding region was inserted  between the PGKp and PGKt via the Xho1 cloning site  This study  pHVXKD3  pHVX2D3 to which a kanMX resistance marker (from pUG6)  was inserted into the Sal1 site of pHVX2D3  This study  pUCMD  pUCTRP1 based plasmid into which the DUR3 expression  cassette (5’‐PGKp‐DUR3‐PGKt‐kanMX‐3’) was PCR blunt  end cloned into the middle of TRP1 via the EcoRV site   This study      Table 5. Genetic cassettes used in the genetic construction and characterization of DUR3 expressing  yeast strains.  Cassette  Description  Reference  DUR1,2  Linear expression cassette containing 5’‐URA3‐PGKp‐ DUR1,2‐PGKt‐URA3‐3’  (Coulon, et al. 2006)  DUR3  Linear expression cassette containing 5’‐TRP1‐PGKp‐ DUR3‐PGKt‐kanMX‐TRP1‐3’  This study      2.2  Culture conditions    E.  coli  DH5α  cells were  used  for molecular  cloning  and  propagation  of  plasmids;  cells were  cultured  according  to  standard  methods  (Ausubel,  et  al.  1995).  Unless  otherwise  indicated,  all  S.  cerevisiae  strains were  cultured  aerobically with  shaking  at 30°C  in  either  liquid  YPD medium  (Difco,  Becton Dickinson  and  Co., USA),  or  on  YPD  +  2%  (w/v)  agar  (Difco,  Becton Dickinson  and  Co., USA)  29    plates. YPD plates supplemented with 300 µg/mL G418 (Sigma, USA) were used to select for positive S.  cerevisiae transformants containing the DUR3 expression cassette.    2.3  Genetic construction of the urea degrading yeasts K7EC‐ and K9EC‐    2.3.1  Co‐transformation of the DUR1,2 cassette and pUT332    S. cerevisiae strains K7 and K9 were co‐transformed with the 9191 bp DUR1,2 cassette (Table 5)  and  pUT332∆ura3  (Gatignol  1987)  combined  at  a  10:1  (DUR1,2  cassette:pUT332) molar  ratio.  Yeast  strains  were  transformed  using  the  lithium  acetate/polyethylene  glycol/ssDNA  method  (Gietz  and  Woods 2002).  Following  transformation,  cells were  left  to  recover  in YEG at 30°C  for 2 hours before  plating  on  YEG  plates  supplemented  with  100  µg/mL  phleomycin  (Invitrogen,  USA).  Plates  were  incubated at 30°C until colonies appeared.    2.3.2  Screening of transformants for the integrated DUR1,2 cassette    Colony PCR  (Ward 1992) was used  as described  to detect  the presence of  the  linear DUR1,2  cassette  integrated  into  the yeast genome at  the URA3  locus. Zymolyase 100 U/mL  (Seikagaku Corp.,  Japan)  was  used  to  lyse  the  cells  (30  µl  zymolyase  solution).  Primers  InDURURA  (5’‐  TGGTGATATGGTTGATTCTGGTGACATA  ‐3’)  and  OutURAb  (5’‐  TTCCAGCCCATATCCAACTTCCAATTTA  ‐3’)  were used to generate an approximately 1500 bp fragment from the 3’ end of the cassette. The primers  are  specific  for  the  inside  and  outside  of  the  integrated  cassette  in  order  to  detect  integration  and  correct orientation. The yeasts 522 and 522EC‐ were used as negative and positive controls, respectively.  PCR was performed with iProofTM High Fidelity DNA polymerase (BioRad, USA) using suggested reagent  concentrations and 1 µl of Zymolyase treated cell suspension as a  template. The PCR program was as  follows: 1. Initial denaturation – 3 min at 98°C. 2. Denaturation – 10 sec at 98°C. 3. Annealing – 30 sec at  58.5°C. 4. Extension – 45 sec at 72°C. 5. Cycle to step 2, 30 times. 6. Final extension – 10 min at 72°C.   Colony PCR reactions were visualized on 0.8% agarose gels stained with SYBR™ Safe (Invitrogen, USA).   After  identification of positive transformants, engineered strains were cultured for ~10 generations on  non‐selective media (YPD) in order to ensure loss of the co‐transforming plasmid.      30    2.3.3  Genetic characterization    2.3.3.1    Southern  blot  analyses.  Southern  blotting  was  used  to  confirm  integration  of  the  DUR1,2  cassette  into  the URA3  locus. Genomic DNA  from  engineered  strains  K7EC‐  and  K9EC‐  as well  as  their  respective parent strains was digested with BglII (Roche, Germany) (Ausubel, et al. 1995), and separated  on  a  0.8%  agarose  gel.  Following  gel  preparation,  transfer  and  fixing  to  a  positively  charged  Nylon  membrane  (Roche  Diagnostics,  Germany)  (Ausubel,  et  al.  1995),  the  blots  were  probed  with  PCR  generated fragments specific for DUR1,2 and URA3. The AlkPhosTM Direct Nucleic Acid Labeling and CDP‐ Star Detection system was used as recommended for probe detection (Amersham Biosciences, England).    The 736 bp DUR1,2 probe was generated by PCR using genomic DNA  from S. cerevisiae strain  522 as a template and the primers DUR1,2probe5 (5’‐TTAGACTGCGTCTCCATCTTTG‐3’) and DUR1,2F (5’‐ TGCTGGCTTTACTGAAGAAGAG‐3’). The 927 bp URA3 probe was generated by PCR using 522 genomic  DNA  and  the  primers  3’URA3  (5'‐TGGGAAGCATATTTGAGAAGATG‐3')  and  OutURAb  (5’‐  TTCCAGCCCATATCCAACTTCCAATTTA ‐3’).    2.3.3.2  Sequence analysis. Genomic DNA isolated from strains K7EC‐ and K9EC‐ was used to amplify two  separate  fragments which  together encompassed  the entire ~10 kb  linear ura3‐PGK1p‐DUR1,2‐PGK1t‐ ura3  cassette.  Primers  5’OUTURA3Cas  (5'‐AACTAATGAGATGGAATCGGTAG‐3')  and  DUR12rev1  (5’‐ TCCTGGAATGCTGTGATGG‐3’) amplified a 7900 bp fragment on the 5’ end of the cassette, while primers  PGK1forDUR  (5’‐TGGTTTAGTTTAGTAGAACCTCGTGAAACTTAC‐3’)  and  OutURAb  (5’‐  TTCCAGCCCATATCCAACTTCCAATTTA ‐3’) amplified a 7100 bp fragment on the 3’ end of the cassette.     Sequencing was performed by the Nucleic Acid Protein Service Unit (NAPS) at The University of  British Columbia using an Applied Biosystems PRISM 377 sequencer and Applied Biosystems BigDye v3.1  sequencing chemistry. Primers and template were supplied to NAPS  in the concentrations specified by  their  sample  submission  requirements.  The  entire  integrated DUR1,2  cassette was  sequenced  via  19  different sequencing reads (Table 6) and later assembled in silico using Accelrys DS Gene v1.1 software.  The assembled  sequences were aligned against previously published  sequences and  those of DUR1,2,  URA3, PGK1p, and PGK1t obtained  from SGD.  If any discrepancies were  found,  the specific read which  gave rise to the discrepancy was repeated in order to identify bona fide mutations.   31      Table 6.  Oligonucleotide primers used in sequencing of the integrated DUR1,2 cassette.   Primer  Primer Name  Sequence (5’Æ3’)  P1  5'OUTURA3Cas  5'‐AACTAATGAGATGGAATCGGTAG‐3'  P2  5'URA3Flank  5'‐AGTATTCTTAACCCAACTGCACAGA‐3'  P3  5'PGK1pro1  5'‐ACAAAATCTTCTTGACAAACGTCACAA3'  P4  5'PGK1pro2  5'‐AATTGATGTTACCCTCATAAAGCACGT‐3'  P5  PGK1forDUR  5'‐TGGTTTAGTTTAGTAGAACCTCGTGAAACTTAC‐3'  P6  DUR12G  5'‐TACCAGAACCTGCTGTATCAG‐3'  P7  DUR12rev6  5'‐TCATCCGCAACTTGTTGCATAG‐3'  P8  DUR12F  5'‐TGCTGGCTTTACTGAAGAAGAG‐3'  P9  DUR12rev5  5'‐TCGGAATAAACTGCAACTGATC‐3'  P10  DUR12E  5'‐ACCTCTGATAATATCTCCCGAAG‐3'  P11  DUR12D  5'‐TTTTGGCCAATGTTGGATCATATTC‐3'  P12  DUR12rev3  5'‐TGTCAACTTGCCAATGGATAAAGTAG‐3'  P13  DUR12C  5'‐TTGTAATGAACCTTCCACTTCTC‐3'  P14  DUR12B  5'‐ACACATGCCAAAGTCTTCGAG3'  P15  DUR12A  5'‐ATTTCCAAAAACGCCCAGAATAC‐3'  P16  DUR12for3end  5'‐TCATCAAGAATACTTGAGATGGATC‐3'  P17  InDURURA  5’‐TGGTGATATGGTTGATTCTGGTGACATA‐3’  P18  3'URA3  5'‐TGGGAAGCATATTTGAGAAGATG‐3'  P19  OutURAb  5’‐TTCCAGCCCATATCCAACTTCCAATTTA‐3’      2.3.3.3   Analysis of DUR1,2 gene expression by qRT‐PCR. Single colonies of parental strains as well as  engineered strains from freshly streaked YPD plates were inoculated into 5 mL YPD and grown overnight  at 30°C on a rotary wheel. Cells were subcultured  into 50 mL YPD (final OD600 = 0.05) and again grown  overnight at 30°C  in a water shaker bath (180 rpm). Cells were harvested by centrifugation (5000 rpm,  4°C, 5 min) and washed once with 50 mL  sterile water. Cell pellets were  resuspended  in 5 mL  sterile  water and OD600 measured. Cell suspensions were used to inoculate sterile 250 mL Schott bottles filled  with 200 mL  filter  sterilized  (0.22 µm, Millipore, USA) Calona Chardonnay  juice  to a  final OD600 = 0.1.  Bottles were  aseptically  sealed with  vapour  locks  (sterilized with  70%  v/v  ethanol)  filled with  sterile  water. Sealed bottles were incubated at 20°C for 24 hours.    Total RNA from 24 hour fermentations was extracted using the hot phenol method (Ausubel, et  al. 1995); RNA was cleaned up post extraction using a total RNeasy Midi Kit (Qiagen), and quantified on  a  Pharmacia Ultrospec  3000 UV/Vis  spectrophotometer.  Clean  total  RNA  (1  µg) was  used  for  cDNA  synthesis  (iScriptTM, BioRad, USA)  according  to  the manufacturer’s  instructions.    iTAQ™  SYBR® Green  32    Supermix with ROX (BioRad, USA) was used  in conjunction with an Applied Biosystems 7500 Real Time  PCR machine  in order  to determine  the  relative  levels of gene expression  in  the  strains  studied. Real  time  primers  for  both  DUR1,2  and  ACT1 were  automatically  optimized  and  designed  using  Applied  Biosystem’s Primer ExpressTM v2.0 software. The DUR1,2 amplification product was amplified using the  primers  DUR12RTfwd  (5’‐CTCTGGTCCAATGGACGCATA  ‐3’)  DUR12RTrev  (5’‐ GATGGATGGACCAGTCAACGTT‐3’).  ACT1  was  amplified  using  the  primers  act1forward  (5’‐ GTTTCCATCCAAGCCGTTTTG‐3’) and act1reverse  (5’‐GCGTAAATTGGAACGACGTGAG‐3’). For each  strain,  DUR1,2 and ACT1 expression was analyzed six times and results were averaged. RQ data were analyzed  using the Applied Biosystems RQ Study software v1.2.2.    2.3.3.4   Global gene expression analysis. Total RNA  from  strains K7 and K7EC‐ was extracted after 24  hours of fermentation (Section 2.3.3.3), via the hot phenol method (Ausubel, et al. 1995), quantified on  a NanoDrop ND‐1000 spectrophotometer and visualized on a 0.8% agarose gel. A ‘GeneChip® One‐Cycle  Target Labeling and Control’ kit (Affymetrix, USA) was used according to manufacturer’s instructions for  synthesis and clean up of cDNA and for synthesis, cleanup and fragmentation of biotinylated cRNA from  10 µg of high quality total RNA. Microarray analyses were done  in duplicate, each with  independently  grown cell cultures.    Four oligonucleotide yeast genome arrays, two per strain, (YGS98; Affymetrix, USA) were used  for  hybridization  of  fragmented  labelled  cRNA.  Preparation  of  hybridization  solution,  hybridization,  washing,  staining,  and  scanning  of  the  microarrays  were  performed  as  per  the  manufacturer’s  instructions  (Eukaryotic Array Gene Chip Expression Analysis and Technical Manual; Affymetrix, USA).  The  EukGE‐WS2v4  fluidics  protocol  of Affymetrix MASv5.0  software was  used  for  array  staining  and  washing  procedures  while  arrays  were  scanned  using  a  G2500A  GeneArray  Scanner  (Agilent  Technologies, USA).    Data  were  analyzed  with MASv5.0  and  DMT  software  (Affymetrix,  USA)  running  on  default  settings  (Affymetrix  Statistical  Algorithm  Reference  Guide).  Statistically  significant  and  reproducible  results were obtained by only  including genes which responded similarly  in all  four cross comparisons  and with change p‐values of ≤0.005 (increasers) or ≥0.995 (decreasers). Reported fold change values are  derived from the average (n=4) of the Signal Log (base 2) Ratio (SLR). Array annotations were linked to  33    their  gene  ontology  (GO)  annotations  using  the  ‘gene_association.sgd.tab’  table  (http://www.yeastgenome.org/gene_list.shtml).    2.3.4  Phenotypic characterization    2.3.4.1   Analysis of fermentation rate  in Chardonnay must. Single colonies of parental strains (K7 and  K9) as well as appropriate engineered strains from freshly streaked YPD plates, were  inoculated  into 5  mL YPD and grown overnight at 30°C on   a rotary wheel. Cells were subcultured  into 50 mL YPD (final  OD600 = 0.05) and again grown overnight at 30°C in a water shaker bath (180 rpm). Cells were harvested  by centrifugation  (5000 rpm, 4°C, 5 min) and washed once with 50 mL sterile water. Cell pellets were  resuspended in 5 mL sterile water and OD600 measured. Cell suspensions were used to inoculate sterile  250 mL Schott bottles filled with 200 mL unfiltered Chardonnay  juice obtained from Calona Vineyards,  Kelowna,  BC,  Canada  to  a  final OD600  =  0.1.  Bottles were  aseptically  sealed with  sterilized  (70%  v/v  ethanol) vapour locks filled with sterile water. Sealed bottles were incubated at 20°C, and weighed daily  to monitor CO2 production. Data were plotted in Excel to generate fermentation profiles.     2.3.4.2  Analysis of fermentation rate in Sake mash. Koji rice was prepared in 400 g batches from short  grain  Japanese Kokoho Rose  (Safeway, Canada). Rice was  rinsed with  tap water at  room  temperature  until  the water  ran  clear;  the  rice was  subsequently  soaked  for  1.5  hours  at  room  temperature  in  enough water to cover the rice. The rice was then drained in a kitchen sieve for 20 min. The rice in the  sieve was then placed over a pot of boiling water and covered with a bamboo steamer lid. The sieve was  arranged  in such a way that the rice was not  in direct contact with boiling water. After steaming until  soft and slightly transparent, the cooked rice was placed  in a stainless steel bowl and cooled to ~30°C.  Upon  reaching  30°C,  the  cooled  rice  was  inoculated  with  1.5  g  of  koji  seeds  (Vision  Brewing,  Washington, USA) mixed with 1 teaspoon of all purpose white flour. The rice was mixed well, covered  with a piece of moist Whatman No. 3 paper, and  the bowl was  sealed with plastic  film. The covered  bowl was incubated, with occasional mixing, at 30°C for 48 hours, or until the rice grains were covered  with fine white fibres and the entire mixture had a cheese‐like aroma. The koji was then transferred to  sterile 500 mL centrifuge bottles and stored at ‐30°C until use.    Single colonies of parental  strains  (K7 and K9) as well as appropriate engineered  strains  from  freshly  streaked YPD plates,   were  inoculated  into 5mL YPD and grown overnight at 30°C on a  rotary  34    wheel. Cells were subcultured into 50 mL YPD (final OD600 = 0.05) and again grown overnight at 30°C in a  water shaker bath (180 rpm). Cells were harvested by centrifugation (5000 rpm, 4°C, 5 min) and washed  once with 50 mL sterile water. Cell pellets were resuspended in 5 mL sterile water and OD600 measured.     The cell suspension was used to inoculate (final OD600 = 0.1) sterile 250 mL Schott bottles filled  with 13 g ‘koji’ rice, 48 g freshly steamed rice (steamed as per ‘koji’ preparation above), and 100 mL of  water  containing  0.125  g/L  citric  acid.  Bottles were  aseptically  sealed with  sterilized  (70%  ethanol)  vapour  locks  filled with  sterile water.  Sealed  bottles were  incubated  at  18°C,  and weighed  daily  to  monitor CO2 production. Data were plotted in Excel to generate fermentation profiles.    2.3.4.3  Analysis of glucose/fructose utilization and ethanol production. Following fermentation, 1 mL  of fresh unheated Sake was transferred to a sterile 1.5 mL microcentrifuge tube and centrifuged for 10  min at max speed (13K RPM) to pellet cells and any particulate. Supernatant (500 μL) was transferred to  a new autosampler screw cap vial (Agilent, USA).    A 20%  (v/v) EtOH standard was made  from 100% EtOH  (Sigma, USA) mixed with sterile MilliQ  water (Millipore, USA), and 1 mL of the standard was transferred to new autosampler screw cap vial. A  standard curve corresponding to 4, 8, 12, 16, and 20% (v/v) EtOH was plotted after duplicate injections  of 2, 4, 6, 8, and 10 µL of the 20% standard as outlined below.    A  3  g/L  glucose/fructose  standard was made  from  D‐Glucose  (Fisher  Scientific, USA)  and  D‐ Fructose  (Fisher  Scientific,  USA) mixed  with  sterile MilliQ  water  (Millipore,  USA),  and  1 mL  of  the  standard was  transferred  to new autosampler  screw cap vial. A  standard curve corresponding  to 0.6,  1.2, 1.8, 2.4, and 3.0 g/L glucose/fructose was plotted after duplicate injections of 2, 4, 6, 8, and 10 µL of  the 20% standard as outlined below.     Samples were analyzed on an Agilent 1100  series  liquid  chromatograph  running Chemstation  Rev A.09.03  [1417]  software  (Agilent Technologies, USA). The LC was  fitted with a Supelcogel C‐610H  main column [column temperature: 50°C, 30 cm x 7.8 mm  ID] (Supelco, USA) that was protected by a  Supelguard C‐610H [5cm x 4.6 mm ID] (Supelco, USA) guard column. A 10 μL sample was run isocratically  with 0.1% (v/v) H3PO4/H20 buffer at a flow rate of 0.75 mL/min. Ethanol was eluted from the column at  35    ~19 min,  fructose was eluted at ~8 min and glucose was eluted at ~9.5 min, and all compounds were  detected by a refractive index detector running in positive mode. The concentration of each compound  was determined automatically by Chemstation software as based on the standard curves.     2.3.5 Functionality analyses    2.3.5.1 Reduction of EC  in Chardonnay wine. Chardonnay wine was produced with K7, K7EC‐, K9, and  K9EC‐ as in Section 2.3.4.1. At the end of fermentation, cells were removed by centrifugation (5000 rpm,  4°C, 5 min), and ~ 50 mL of wine was decanted into sterile 50 mL Schott bottles. Bottles were incubated  in a 70°C water bath for exactly 48 hours to maximize EC production, and then stored at 4°C until GC/MS  analysis (Section 2.3.5.3).    2.3.5.2 Reduction of EC in Sake wine. Sake wine was produced with K7, K7EC‐, K9, and K9EC‐ as in Section  2.3.4.2. At  the end of  fermentation, cells and  rice were  removed by centrifugation  (5000  rpm, 4°C, 5  min), and ~ 50 mL of wine was decanted  into sterile 50 mL Schott bottles. Bottles were  incubated  in a  70°C water bath  for exactly 48 hours  to maximize EC production, and  then stored at 4°C until GC/MS  analysis (Section 2.3.5.3).    2.3.5.3   Quantification of EC  in wine by solid phase microextraction and GC/MS. A 10 mL sample of  heated (70°C  ‐ 48 hours) wine was pipetted  into a 20 mL sample vial, to which a magnetic stirring bar  and 3 g of NaCl were added. The vial was capped with a PTFE/silicone septum, placed on a stirrer at  22°C, and allowed to equilibrate, while stirring, for 15 min. A Solid Phase Microextraction (SPME) fibre  [65 µm Carbowax/Divinylbenzene  (CW/DVB)] was  conditioned  at  250°C  for  30 min before use. After  sample equilibration, the fibre was  inserted  into the head space. After 30 min, the fibre was removed  from the sample vial and inserted into the injection port for 15 min. A blank run was performed before  each sample run. Quantification was done as follows: an ethyl carbamate (Sigma‐Aldrich) standard stock  solution (0.1 mg/mL) was prepared in distilled H20 containing 12% (v/v) ethanol and 1 mM tartaric acid  at pH 3.1. Calibration standards were prepared with ethyl carbamate concentrations of 5, 10, 20, 40, 90  µg/L. The standard solutions were stored in the refrigerator at 4°C.    Ethyl carbamate in wine was quantified using an Agilent 6890N GC interfaced to a 5973N Mass  Selective Detector. A 60 m x 0.25 mm  ID, 0.25 µm thickness DBWAX  fused silica open tubular column  36    (J&W  Scientific,  Folsom,  CA,  USA) was  employed.  The  carrier  gas was  ultra  high  purity  helium  at  a  constant  flow  of  36  cm/s.  The  injector  and  transfer  line  temperature  was  set  at  250°C.  The  oven  temperature was initially set at 70°C for 2 min then raised to 180°C at 8°C /min and held for 3 min. The  temperature was then programmed to increase by 20°C /min to 220°C where it was held for 15 min. The  total run time was 35.75 min. The  injection mode was splitless for 5 min (purge flow: 5 mL/min, purge  time:  5  min).  The MS  was  operated  in  Selected  Ion Monitoring  (SIM)  mode  with  electron  impact  ionization; MS quad  temperature 150°C  and MS  source  temperature 230°C. The  solvent delay was 8  min. Specific  ions 44, 62, 74, 89 were monitored with a dwell time of 100 msec. Mass 62 was used for  quantification  against  the  mass  spectrum  of  the  authentic  ethyl  carbamate  standard.  Wines  were  analyzed by three separate injections and the data were averaged.     2.4  Genetic construction of the urea importing yeasts K7D3, K7EC‐D3, 522D3, and 522EC‐D3    2.4.1  Construction of the DUR3 linear cassette    In order  to  express DUR3  constitutively,  a  cassette  similar  to  the DUR1,2  cassette previously  made  by  our  group  (Coulon,  et  al.  2006)  was  constructed.  All  cloning  steps  and  reactions,  unless  otherwise  stated,  were  performed  according  to  standard  molecular  biology  standards  methods  (Ausubel, et al. 1995). All PCR reactions, unless otherwise indicated, were performed using iProofTM High  Fidelity DNA polymerase (BioRad, USA) as per the manufacturer’s instructions.    2.4.1.1   Construction of pHVX2D3.  In order to place DUR3 under the control of the constitutive PGK1  promoter and terminator signals, the DUR3 ORF was cloned into pHVX2 (Volschenk, et al. 1997) (Figure  10). The DUR3 ORF was amplified from 522 genomic DNA using the following primers which contained  Xho1 restriction enzyme sites built into their 5’ ends:   1. DUR3forXho1 (5’‐AAAACTCGAGATGGGAGAATTTAAACCTCCGCTAC‐3’)   2. DUR3revXho1 (5’‐AAAACTCGAGCTAAATTATTTCATCAACTTGTCCGAAATGTG‐3’).    Following PCR, 0.8% agarose gel visualization, and PCR cleanup (Qiagen, USA – PCR Purification  Kit), both the PCR product (insert) and pHVX2 (vector) were digested with Xho1 (Roche, Germany). After  the digested vector was treated with SAP (Fermentas, USA) to prevent re‐circularization, the insert and  linearized‐SAP  treated  vector were  ligated overnight  at  22°C  (T4 DNA  Ligase  –  Fermentas, USA);  the  37    ligation mixture  (5  μL) was  used  to  transform  DH5α™  competent  cells  (Invitrogen,  USA)  that were  subsequently  grown  on  100  µg/mL  Ampicillin  (Fisher,  USA)  supplemented  LB  (Difco,  USA)  plates.  Plasmids from a random selection of transformant colonies were harvested (Qiagen, USA – QIAprep Spin  Miniprep kit) and digested by EcoR1 (Roche, Germany); PCR, using inside‐outside primers, was done to  identify plasmids with the desired insert.      Figure 10. Schematic  representation of cloning  strategy  for creation of pHVX2D3. The DUR3 ORF was  PCR amplified from 522 genomic DNA and ligated into the Xho1 site of pHVX2.      2.4.1.2  Construction of pHVXKD3. A kanMX marker was obtained from pUG6 (Guldener, et al. 1996) via  double digestion with Xho1 and Sal1  (Fermentas, USA). Following digestion, the 1500 bp kanMX band  was gel purified (Qiagen, USA – Gel Extraction Kit) and ligated into the Sal1 site of linearized‐SAP treated  pHVX2D3. The ligation mixture (5 μL) was used to transform DH5α™ competent cells which were grown  on  LB‐Ampicillin  (100  µg/mL).  Recombinant  plasmids  (Figure  11)  were  identified  by  HindIII  (Roche,  Germany) digestion of plasmids isolated from 24 randomly chosen colonies.    38      Figure 11. Schematic  representation of  cloning  strategy  for  creation of pHVXKD3. The kanMX marker  was obtained from Xho1/Sal1 digestion of pUG6 and ligated into the Sal1 site of pHVXKD3.    2.4.1.3 Construction of pUCTRP1. The TRP1 coding  region was PCR amplified  from 522 genomic DNA  using TRP1 specific primers, each containing BamH1  (bold) and  then Apa1  (underline) sites at  their 5’  ends: BamH1Apa1TRP1ORFfwd (5’‐AAAAAAGGATCCAAAAAAGGGCCCATGTCTGTTATTAATTTCACAGG‐3’);  BamH1Apa1TRP1ORFrev (5’‐AAAAAAGGATCCAAAAAAGGGCCCCTATTTCTTAGCATTTTTGACG‐3’).     Following amplification, cleanup, and quantification, the ~750 bp fragment was ligated into the  BamH1 (Roche, Germany) site of linearized‐SAP treated pUC18 (Figure 12). Recombinant plasmids were  identified primarily through blue/white screening (growth on LB‐Ampicillin supplemented with 50 μg/mL  Xgal) and subsequently confirmed through HindIII/EcoR1 digestion.      Figure 12. Schematic representation of cloning strategy for creation of pUCTRP1. The TRP1 ORF was PCR  amplified from 522 genomic DNA and ligated into the BamH1 site of pUC18.    39    2.4.1.4  Construction of pUCMD. The PGK1p‐DUR3‐PGK1t‐kanMX cassette located within pHVXKD3 was  amplified from pHVXKD3 plasmid DNA using cassette specific primers: 1. pHVXKfwdlong (5’‐ CTGGCACG  ACAGGTTTCCCGACTGGAAAGCGGGCAGTGAG‐3’); pHVXKrevlong (5’‐ CTGGCGAAAGGGGGATGTGCTGCAA  GGCGATTAAGTTGGG‐3’). Following amplification, cleanup, and quantification,  the ~6500 bp blunt end  PCR generated fragment was treated with polynucleotide kinase (New England Biolabs, USA) in order to  facilitate  ligation  (O/N  at  22°C)  into  the blunt  EcoRV  (Fermentas, USA)  site  of  linearized‐SAP  treated  pUCTRP1.    Recombinant plasmids (Figure 13) were  initially  identified using E‐lyse analysis (Eckhardt 1978)  and  later confirmed via Apa1  (Stratagene, USA) /Sal1 digestion. Briefly, E‐lyse efficiently screens  large  numbers  of  colonies  for  the  presence  of  plasmid DNA  by  lysing  the  colonies within  the wells  of  an  agarose gel, followed by electrophoresis (Eckhardt 1978). More specifically, after patching onto selective  media,  small aliquots of  colonies were  suspended  in 5 µL TBE buffer and  then mixed with 10 µL SRL  buffer  (25%  v/v  sucrose,  50  µg/mL  RNaseA,  1  mg/mL  lysozyme).  After  mixing  by  pipetting,  cell  suspensions were  loaded  into  the wells of  a  0.2%  (w/v)  SDS  ‐  0.8%  (w/v)  agarose  gel. After  the  cell  suspension in the wells had become clear indicating cell lysis (~ 30 min), the DNA was electrophoresed  at 20 V  for 45 min,  then at 80 V  for 45 min. Finally  the gel was stained as  required with SYBR™ Safe  (Invitrogen, USA).    40      Figure 13. Schematic representation of cloning strategy for creation of pUCMD. The linear PGK1p‐DUR3‐ PGK1t‐kanMX construction was PCR amplified from pHVXKD3 and blunt end cloned into the EcoRV site  of pUCTRP1 to yield pUCMD.      2.4.2  Sequence analysis of the DUR3 cassette in pUCMD    Plasmid pUCMD (Figure 14) was isolated from E. coli (Qiagen, USA – QIAprep Spin Miniprep kit)  and  used  directly  as  a  sequencing  template.  Sequencing was  performed  by  the Nucleic Acid  Protein  Service  Unit  (NAPS)  at  The  University  of  British  Columbia  using  an  Applied  Biosystems  PRISM  377  sequencer  and  Applied  Biosystems  BigDye  v3.1  sequencing  chemistry.  Primers  and  template  were  supplied to NAPS  in the concentrations specified by their sample submission requirements. The entire  DUR3  cassette,  beginning  from  the  5’  TRP1  flanking  sequence  and  ending with  the  3’  TRP1  flanking  sequence, was sequenced via 16 different sequencing reads (Table 6) and later assembled in silico using  Accelrys DS Gene v1.1  software. The assembled  sequences were aligned against previously published  sequences and  those of DUR3, TRP1, PGK1p, and PGK1t obtained  from SGD.  If any discrepancies were  found, the specific read which gave rise to the discrepancy was repeated in order to identify bona fide  mutations.   41      Table 7. Oligonucleotide primers used in sequencing of DUR3 cassette in pUCMD.   Primer  Primer Name  Sequence (5’Æ3’)  P1  BamH1TRP1Apa1Fwd  5’‐AAAAAAGGATCCAAAAAAGGGCCCATGTCTGTTATTAATTTCACAGG‐3’  P2  pHVXKlongfwd  5’‐CTGGCACGACAGGTTTCCCGACTGGAAAGCGGGCAGTGAG‐3’  P3  pHVXKfwd  5’‐CTGGCACGACAGGTTTCCCGACTGG‐3’  P4  pDUR3tfwd  5’‐TTTCCGCGGAGCTTTCTAACTGATCTATCC‐3’  P5  DUR3Xho1fwd  5’‐AAAACTCGAGATGGGAGAATTTAAACCTCCGCTAC‐3’  P6  DUR3Xho1rev  5’‐AAAACTCGAGCTAAATTATTTCATCAACTTGTCCGAAATGTG‐3’  P7  pDUR3trev  5’‐TTTCCGCGGTGCGGTGTGAAATACC‐3’  P8  kanMXORFrev  5’‐TTAGAAAAACTCACTGAGCATCAAATGAAACTGC‐3’  P9  kanMXORFfwd  5’‐ATGGGTAAGGAAAAGACTCACGTTTCGAGG‐3’  P10  pHVXKlongrev  5’‐CTGGCGAAAGGGGGATGTGCTGCAAGGCGATTAAGTTGGG‐3’  P11  BamH1TRP1Apa1rev  5’‐AAAAAAGGATCCAAAAAAGGGCCCCTATTTCTTAGCATTTTTGACG‐3’  P12  DUR3RTfwd  5’‐GATCGGCCATGGTTGCTACTT‐3’  P13  RevPGKtPst1  5’‐TTTTCTGCAGAAGCTTTAACGAACGCAGAATT‐3’  P14  PGKpro1  5'‐ACAAAATCTTCTTGACAAACGTCACAA3'  P15  PGKpro2  5'‐AATTGATGTTACCCTCATAAAGCACGT‐3'  P16  PGKforDUR  5'‐TGGTTTAGTTTAGTAGAACCTCGTGAAACTTAC‐3'  42    Figure 14. Schematic representation of pUCMD. The DUR3 linear cassette stretches between Apa1 sites encompassing 5’1/2TRP1‐PGK1p‐ DUR3‐PGK1t‐kanMXp‐kanMX‐kanMXt‐3’1/2TRP1.      5’ ½ TRP1  PGK1 Promoter  PGK1 Terminator  DUR3  kanMX Promoter  kanMX Terminator  kanMX  3’ ½ TRP1  AmpR  rep (pMB1)  pUCMD (9221bp)     43    2.4.3  Transformation of the linear DUR3 cassette into S. cerevisiae and selection of transformants    The 6536 bp DUR3 cassette was cut from pUCMD using Apa1 (Stratagene, USA) (Ausubel, et al.  1995) and visualized on a 0.8% agarose gel. From the gel, the expected 6536 bp band was resolved and  extracted  (Qiagen,  USA  –  Gel  extraction  kit).  After  extraction,  clean  up,  and  quantification  using  a  Nanodrop  ND‐1000  spectrophotometer  (Nanodrop,  USA),  250  ng  of  linear  cassette  was  used  to  transform  S.  cerevisiae  strains  K7,  K7EC‐,  522,  and  522EC‐.  Yeast  strains  were  transformed  using  the  lithium acetate/polyethylene glycol/ssDNA method (Gietz and Woods 2002). Following transformation,  cells were left to recover in YPD at 30°C for 2 hours before plating on to YPD plates supplemented with  300 µg/mL G418 (Sigma, USA). Plates were incubated at 30°C until colonies appeared.     2.4.3.1   Confirmation of  integration via colony PCR. Colony PCR was used as previously described  to  detect  the presence of  the  linear DUR3  cassette  integrated  into  the yeast genome at  the TRP1  locus  (Ward 1992). Zymolyase 100 U/mL (Seikagaku Corp., Japan) was used to lyse the cells (30 µl zymolyase  solution).  Primers  kanMXORFfwd  (5’‐ATGGGTAAGGAAAAGACTCACGTTTCGAGG‐3’)  and  kanMXORFrev  (5’‐TTAGAAAAACTCATCGAGCATCAAATGAAACTGC‐3’) were used  to generate an 809 bp  fragment  from  the 3’ end of the cassette. 522 genomic DNA was used as a negative control and pUCMD as a positive  control. PCR was performed with iProofTM High Fidelity DNA polymerase (BioRad, USA) using suggested  reagent concentrations and 1 µl of Zymolyase  treated cell supernatant as  template. The PCR program  was as follows: 1. Initial denaturation – 3 min at 98°C. 2. Denaturation – 10 sec at 98°C. 3. Annealing –  20 sec at 62.4°C. 4. Extension – 30 sec at 72°C. 5. Cycle to step 2, 30 times. 6. Final extension – 10 min at  72°C.   Colony PCR reactions were visualized on 0.8% agarose gels stained with SYBR™ Safe (Invitrogen,  USA).     2.4.4  Genetic characterization    2.4.4.1  Southern blot analyses. Southern blotting was used to confirm integration of the DUR3 cassette  into  the TRP1  locus. Genomic DNA  from engineered  strains K7EC‐, K7D3, K7EC‐D3, 522D3, and 522EC‐D3, as  well as their respective parent strains, was digested with EcoRI (Roche, Germany) (Ausubel, et al. 1995),  and  separated  on  a  0.8%  agarose  gel.  Following  gel  preparation,  transfer  and  fixing  to  a  positively  charged Nylon membrane  (Roche Diagnostics, Germany)  (Ausubel, et al. 1995), the blots were probed  with PCR generated fragments specific for DUR3 and TRP1. The AlkPhosTM Direct Nucleic Acid Labeling      44    and CDP‐Star Detection system was used as recommended for probe detection (Amersham Biosciences,  England).    The 661 bp DUR3 probe was generated using genomic DNA  from S. cerevisiae strain 522 as a  template  and  the  primers  DUR3probefwd  (5’‐  CAGCAGAAGAATTCACCACCGCCGGTAGATC‐3’)  and  DUR3proberev  (5’‐  CAATCAGGTTAATAATTAATAAAATACCAGCGG‐3’).  The  461  bp  TRP1  probe  was  generated  using  genomic  DNA  from  522  as  a  template  and  the  primers  TRP1probefwd  (5’‐  TTAATTTCACAGGTAGTTCTGGTCCATTGG‐3’) and TRP1proberev (5’‐ CAATCCAAAAGTTCACCTGTCCCACCT  GCTTCTG‐3’).     2.4.4.2  Analysis of gene expression by northern blotting.  Fermentations with K7, K7EC‐ K7D3, and K7EC‐D3  in  filter sterilized Calona Chardonnay must were conducted as  in Section 2.3.3.3. After 24 hours, cells  were harvested by  centrifugation  (5000  rpm, 4°C, 4 min),  snap  frozen  in  liquid nitrogen  (3 min), and  stored at ‐80°C until RNA extraction. Total RNA was extracted using a hot phenol method (Ausubel, et al.  1995), quantified on a NanoDrop ND‐1000 spectrophotometer and visualized on a 0.8% agarose gel.    Northern blot analysis was performed as described  (Ausubel, et al. 1995). Briefly, 30  μg  total  RNA  was  separated  on  a  1%  agarose‐formaldehyde  denaturing  gel  and  transferred  to  a  positively  charged  Nylon  membrane  (Roche  Diagnostics,  Germany).  Blots  were  probed  with  PCR  generated  fragments specific  for DUR3 and the  loading control HHF1. The AlkPhosTM Direct Nucleic Acid Labeling  and CDP‐Star Detection system was used as recommended for probe detection (Amersham Biosciences,  England).     The 661 bp DUR3 probe used for northern blotting was the same as the probe used for Southern  blotting (Section 2.4.4.1). The ~500 bp HHF1 probe was supplied by another member of our lab (Coulon,  et al. 2006).    2.4.4.3  Analysis of DUR3 gene expression by qRT‐PCR. Gene expression analysis of K7, K7EC‐ K7D3, and  K7EC‐D3  was  performed  as  described  in  Section  2.3.3.3  using  total  RNA  from  24  hour  fermentations  (Section 2.4.4.2).  The DUR3  real  time PCR product was  generated using  the primers DUR3RTfwd  (5’‐ GATCGGCCATGGTTGCTACTT‐3’) and DUR3RTrev (5’‐GCGATAGTGTTCATCCCGGTT‐3’).       45    2.4.4.4   Global gene expression analysis. Following 24 hour fermentations (Section 2.4.4.2), total RNA  from strains K7 and K7D3 was used for transcriptome analysis as described in Section 2.3.3.4.    2.4.5  Analysis of urea uptake using 14C‐urea    In order to assess the effect of DUR3 constitutive expression on urea uptake activity, a 14C‐urea  uptake assay was performed as previously described  (Cooper and Sumrada 1975). Briefly, appropriate  strains were grown (30°C – 180 RPM)  in minimal media (1.7 g/L YNB w/o amino acids w/o ammonium  sulfate, 20 g/L glucose, 1 g/L ammonium sulfate or 1 g/L L‐proline) to approximately 1x107 cells/mL. An  11 mL sample of cell culture was transferred to a 250 mL Erlenmeyer flask containing 80 μL of 36.6 mM  14C‐urea (Sigma, USA – 6.8 mCi/mmol). Cells were then cultured (30°C – 180 RPM) for 20 min and 1 mL  samples were  taken every 2 min. The 1 mL  samples were applied  to 0.22  μm nylon  filters  (Millipore,  USA) and washed with 25 mL aliquots of cold minimal media  to which 10 mM urea  (Fisher, USA) had  been added. Filters were placed  in scintillation vials, filled with scintillation fluid  (Fisher, USA) and  left  overnight to equilibrate. Samples were then counted  in a recently calibrated Beckman LS6000IC  liquid  scintillation counter using  the counter’s  factory  ‘14C quench’ mode. DPM values were then converted  into nano‐mole urea transported (Cooper and Sumrada 1975) and plotted against time in Excel.    2.4.6  Phenotypic characterization    2.4.6.1    Analysis  of  fermentation  rate  in  Chardonnay  must.  Fermentations  of  unfiltered  Calona  Chardonnay  must  with  K7,  K7EC‐,  K7D3,  K7EC‐D3,  522,  522EC‐,  522D3,  and  522EC‐D3  were  performed  as  described in Section 2.3.4.1.    2.4.6.2   Analysis of  fermentation  rate  in Sake mash. Fermentations of Sake  rice mash with K7, K7EC‐,  K7D3, K7EC‐D3, 522, 522EC‐, 522D3, and 522EC‐D3 were performed as described in Section 2.3.4.2.    2.4.6.3   Analysis for ethanol content. Ethanol production by K7, K7EC‐, K7D3, K7EC‐D3, 522, 522EC‐, 522D3,  and 522EC‐D3 in Chardonnay and Sake wine was quantified as described in Section 2.3.4.3.            46    2.4.7  Functionality of metabolically enhanced yeasts    2.4.7.1  Reduction of EC in Chardonnay wine. Chardonnay wine was produced with K7, K7EC‐, K7D3, K7EC‐ D3, 522, 522EC‐, 522D3, and 522EC‐D3 as described in 2.3.4.1. Quantification of EC reduction was performed  as described in Section 2.3.5.1.    2.4.7.2  Reduction of EC in Sake wine. Sake wine was produced with K7, K7EC‐, K7D3, K7EC‐D3, 522, 522EC‐,  522D3, and 522EC‐D3 as described in 2.3.4.2. Quantification of EC reduction was performed as described in  Section 2.3.5.2.    2.5 Statistical analyses    Two  factor  ANOVA  analyses  were  used  to  evaluate  the  variations  in  glucose,  fructose,  and  ethanol measured  in Chardonnay and  Sake wine produced by parental and engineered  yeast  strains.  Fisher’s LSD  (Least Significant Difference) test was used after ANOVA to determine which means were  statistically  significant  (p<0.05).  All  statistical  calculations were  performed  in  Excel  2007  (Microsoft,  USA).                    47    URA3  PGKp  URA3 PGKt DUR1,2  Srf1 ½ site  Srf1 ½ site  3  RESULTS    3.1  Constitutive expression of DUR1,2 in Sake yeast strains K7 and K9    3.1.1  Integration of the linear DUR1,2 cassette into the genomes of Sake yeast strains K7 and K9    In order  to constitutively express DUR1,2,  the Sake yeast strains K7 and K9 were  transformed  with  the  linear DUR1,2  cassette  (Figure 15). After  colony PCR  screening of approximately 1500  yeast  transformants for the integration of the DUR1,2 cassette, two metabolically engineered strains, K7EC‐ and  K9EC‐, were obtained. The designation  ‘EC‐‘ denotes  integration of  the  linear DUR1,2 cassette  into  the  URA3 locus.      3.1.2  Genetic characterization of K7EC‐ and K9EC‐    3.1.2.1  Correct integration of the DUR1,2 linear cassette into the genomes of K7EC‐ and K9EC‐. In order  to  confirm  the  correct  integration of  the DUR1,2  cassette, a Southern blot was performed using PCR  generated DUR1,2 and URA3 probes and the BglII digested genomic DNA of K7EC‐ and K9EC‐.       When probed with DUR1,2, two signals were detected for K7EC‐ and K9EC‐ corresponding to 5.0  kb and 9.0 kb DNA fragments, and one signal was detected for K7 and K9 corresponding to a 5.0 kb DNA  fragment (Figure 16a). The 5.0 kb fragment matches the expected fragment size for the native DUR1,2  locus while  the 9.0 kb  fragment matches  the expected  fragment  size  for  the presence of  the DUR1,2  cassette integrated into the URA3 locus (Figure 16b).      Figure 15. Schematic representation of the linear DUR1,2 cassette.        48    Probing with URA3 gave rise to two signals for K7EC‐ and K9EC‐, corresponding to 3.8 kb and 4.4 kb  DNA fragments, and one signal for K7 and K9 corresponding to a 4.4 kb fragment (Figure 17a). The 3.8  kb  fragment matches  the  expected  fragment  size  for  the  recombinant  URA3  locus,  and  the  4.4  kb  fragment matches the expected fragment size for a non‐disrupted URA3 locus (Figure 17b).         Figure  16.  Integration  of  the  DUR1,2  cassette  into  the  URA3  locus  of  K7EC‐  and  K9EC‐ was  confirmed  by  Southern  blot  analyses  using  a  DUR1,2  probe  (a).  All  faint  bands  that  do  not  correspond to a size in the schematic representation were deemed to be non‐specific. For each  sample, the agarose gel is shown on the left while the exposed film is on the right. b) Schematic  representation  of  the  signals  expected  during  Southern  blot  analyses  (DUR1,2  probe)  of  recombinant yeasts containing the recombinant DUR1,2 cassette integrated into the URA3 locus.  The DUR1,2 probe is illustrated with blue hatched lines (b).    BglII DUR1,2PGK1p PGK1t ura3ura3 BglII 9085 bp DUR1,2 BglII 5084 bp BglII a)  b)      49                Figure 17. Disruption of the URA3 locus by integration of the DUR1,2 cassette in K7EC‐ and K9EC‐   was confirmed by Southern blot analyses using a URA3 probe  (a). All  faint bands  that do not  correspond to a size in the schematic representation were deemed to be non‐specific. For each  sample, the agarose gel is shown on the left while the exposed film is on the right. b) Schematic  representation  of  the  signals  expected  during  Southern  blot  analysis  (URA3  probe)  of  recombinant  yeasts  containing  the  recombinant  DUR1,2  cassette  integrated  into  the  URA3  locus. The URA3 probe is illustrated with green hatched lines (b).  BglII DUR1,2PGK1p PGK1t ura3ura3 BglII 3757 bp URA3 BglII 4455 bp BglII a)  b)      50    3.1.2.2   Sake strains K7EC‐ and K9EC‐ do not contain the bla and Tn5ble antibiotic resistance markers.  After transformation, K7EC‐ and K9EC‐ were successively sub‐cultured on a non‐selective medium in order  to eliminate pUT332, whose only purpose was to facilitate early screening for transformants. A Southern  blot using probes  specific  for bla  (Ampicillin  resistance)  and  Tn5ble  (Phleomycin  resistance)  revealed  that these genes were absent from K7EC‐ and K9EC‐ , as well as the parental strains K7 and K9 (Figure 18).     Figure 18. The genetically engineered strains K7EC‐ and K9EC‐ do not contain the bla and Tn5ble antibiotic  resistance markers. The plasmid pUT332 was used as a positive control for both bla and Tn5ble. For each  sample, the agarose gel is shown on the left while the exposed film is on the right.      51    3.1.2.3  Sequence of the DUR1,2 cassette integrated into the genomes of K7EC‐ and K9EC‐. To verify the  sequence of the DUR1,2 cassette  integrated  into the URA3  locus, single strand DNA sequencing of the  cassette in K7EC‐ and K9EC‐ was completed. In silico sequence assembly and subsequent analysis revealed  one nucleotide in the cassette sequence of K9EC‐ that did not match the previously reported recombinant  DUR1,2  sequences and/or SGD’s S288C DUR1,2  sequence  (Table 8). The C  to T  switch  (theoretical  to  sequenced data) at nucleotide position 821 is located within the 5’URA3 flanking region of the cassette  and is likely due to a genetic polymorphism between the Sake strain K9 and the laboratory strain S288C.  Comparison of  the  recombinant DUR1,2 ORF sequence  in K7EC‐ and K9EC‐ and  that of SGD  revealed no  nucleotide  changes  or  amino  acid  substitutions.  A  detailed  description  of  the  DNA  sequences  that  comprise the DUR1,2 cassette is given in Table 9.    Table 8. Discrepancies between the integrated DUR1,2 cassette of K9EC‐ and published sequences.    Nucleotide  position  Description  821  Difference in the 5’ region of the URA3 open reading frame of K9EC‐ CÆT     Table 9. Detailed description of the DNA sequences that comprise the DUR1,2 cassette.  Nucleotide  position  Description  1‐4  5’ SRF1 ½ site  5‐950  URA3 sequence  5‐508  5’ non coding sequence  509‐950  5’ part of URA3 ORF  951‐2445  PGK1 promoter  2446‐7953  DUR1,2 ORF  7954‐8235  PGK1 terminator  8236‐9187  URA3 sequence  8236‐8640  3’ part of URA3 ORF  8641‐9187  3’ non coding sequence  9188‐9191  3’ SRF1 ½ site    In silico analysis of  the  integrated DUR1,2 cassette  revealed  that  two new ORFs were created  during  construction  of  the  DUR1,2  cassette;  these  ORFs  were  composed  entirely  of  S.  cerevisiae  sequences (Figure 19). Novel ORF1 (447 bp) is located at nucleotide position 509‐955 while novel ORF2  (792 bp) is located at position 767‐1558.      52        3.1.2.4   Confirmation of constitutive expression of DUR1,2  in K7EC‐ and K9EC‐ by qRT‐PCR. Total RNA  from  yeast  cells  in  24  hour  fermentations  of  Chardonnay must  was  used  to  confirm  and  quantify  constitutive expression of DUR1,2 in K7EC‐ and K9EC‐. Integration of the DUR1,2 cassette in K7EC‐ and K9EC‐  up  regulated DUR1,2  expression  by  9.13‐fold  and  12.77‐fold,  respectively,  compared  to  the  parental  strains K7 and K9 (Figure 20). Expression of DUR1,2 in K7EC‐ and K9EC‐ was detected in non‐inducing (NCR)  conditions  indicating  that  the  PGK1  promoter  and  terminator  signals  are  effective  at  overcoming  repression by NCR during fermentation.   Figure 19. A  schematic  representation of new ORFs of more  than 100  codons generated during  construction of the DUR1,2 cassette. Two new ORFs, entirely composed of S. cerevisiae sequences,  were created.     Novel ORF2 (792 bp)  Novel ORF1  (447 bp)  5’ URA3  PGK1p  DUR1,2 5’ SRF1 ½ site  PGK1t  3’ URA3     53    1.00 9.13 1.00 12.77 0 2 4 6 8 10 12 14 16 18 K7 K7EC- K9 K9EC- Strain R el at iv e qu an tif ic at io n (fo ld ) R el at iv e qu an tif ic at io n (fo ld )   Figure 20. Gene expression analysis  (qRT‐PCR) of K7, K7EC‐, K9, and K9EC‐  indicates  functionality of  the  DUR1,2 cassette and constitutive expression of DUR1,2 in non‐inducing (NCR) conditions. Total RNA was  extracted  from  yeast  cells  harvested  after  24  hour  fermentation  (20°C)  in  filter  sterilized  Calona  Chardonnay must  that  was  inoculated  to  a  final  OD600  =  0.1.  Total  RNA  was  subsequently  reverse  transcribed,  and  the  resultant  cDNA was  amplified  in  the presence of  SYBR  green dye. DUR1,2 gene  expression was standardized to ACT1 expression and data for strains K7EC‐ and K9EC‐ were calibrated to  their  respective  parental  strain  K7  and  K9.  Fermentations were  conducted  in  triplicate  and  the  data  averaged; error bars represent 95% confidence intervals.     3.1.2.5   Effect of the  integrated DUR1,2 cassette on the transcriptomes of K7EC‐ and K9EC‐. Total RNA  from  yeast  cells  in  24  hour  fermentations  of  Chardonnay must was  used  analyze  the  impact  of  the  integrated DUR1,2  cassette on  global  gene  expression  in K7EC‐. Reported  changes  in  gene  expression  were cut off at a minimum 4‐fold change in expression (SLRAvg ≥ 2 or SLRAvg ≤ ‐2) to ensure elimination of  experimental noise inherent in microarray analysis; this cut‐off is supported by the previously published  statistical examination of a wine yeast’s transcriptome during fermentation (Marks, et al. 2008).    Integration of  the DUR1,2  cassette  into  the URA3  locus of K7EC‐ had  a minimal effect on  the  yeast’s transcriptome. DUR1,2 was upregulated by 6.35‐fold in K7EC‐; URA3 was downregulated by 2.35‐     54    fold but was not included in Table 10 as it fell below the 4‐fold cut off. Besides DUR1,2, two genes were  upregulated greater than 4‐fold in K7EC‐ (Table 10); seven genes were downregulated more than 4‐fold in  K7EC‐.  No  metabolic  pathways  were  affected  by  the  presence  of  the  integrated  DUR1,2  cassette;  however,  integration  of  the  DUR1,2  cassette  downregulated  three  unrelated  genes  involved  in  meiosis/sporulation  (RME1, SSP1, SDS3)  indicating a possible negative effect on  sporulation efficiency  (Table 10).      Table  10.  Effect  of  the  integrated  DUR1,2  cassette  in  the  genome  of  K7  on  global  gene  expression  patterns in S. cerevisiae K7EC‐ (≥ 4‐fold change). Reported changes are relative to the parental strain K7.  Total RNA  from K7 and K7EC‐, harvested at 24 hours  into  fermentation of  filter  sterilized Chardonnay  must, were used  for hybridization  to microarray. Fermentations were conducted  in duplicate and  the  data were averaged (p≤0.005).     Genes expressed at higher levels in K7EC- Fold Change Gene Symbol Biological Process 6.60 RTG1 Transcription factor (bHLH) involved in interorganelle communication 6.35 DUR1,2 Urea amidolyase 4.23 HAC1 bZIP (basic-leucine zipper) protein involved in unfolded protein response Genes expressed at lower levels in K7EC- Fold Change Gene Symbol Biological Process -5.64 RME1 Zinc finger protein involved in control of meiosis -4.97 SEO1 Permease involved in methionine transport -4.67 MF(Alpha)1 Mating factor alpha -4.64 SSP1 Protein involved in the control of meiotic nuclear divisions and spore formation -4.21 SDS3 Protein involved in deactylase complex and transcriptional silencing during sporulation -4.16 CBP1 Protein required for Cytochrome B mRNA stability or 5’ processing -4.12 RUD3 Protein involved in organization of Golgi   3.1.3  Phenotypic characterization of K7EC‐ and K9EC‐    3.1.3.1  Fermentation rate of K7EC‐ and K9EC‐ in Chardonnay must. Fermentation profiles of the parental  and metabolically  engineered  Sake  yeasts  are  shown  in  Figures  21a,b. As  is  common  in  grape must,  fermentations were  robust and  rapid  (~400 hours). The  fermentation profiles  shown  in Figures 21a,b  indicate substantial equivalence amongst the parent and engineered strains as far as fermentation rate  is concerned.        55      0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 0 100 200 300 400 500 T im e   (Hours) W ei g h t  lo ss  ( g ) K 7 K 7E C ‐   0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 0 100 200 300 400 500 T im e   (Hours) W ei g h t  lo ss  ( g ) K 9 K 9E C ‐   Figure  21.  Fermentation  profiles  (weight  loss)  of  parental  and  DUR1,2  engineered  Sake  strains  in  Chardonnay wine. Chardonnay wine was produced by inoculating Sake yeast strains (a) K7 and K7EC‐ and  (b)  K9  and  K9EC‐  into  unfiltered  Calona  Chardonnay  must  (final  OD600  =  0.1).  Fermentations  were  incubated to completion (~400 hours) at 20°C. Fermentations were conducted in triplicate and the data  were averaged; error bars indicate one standard deviation.      3.1.3.2   Fermentation rate of K7EC‐ and K9EC‐  in Sake mash. Fermentation profiles of  the parental and  metabolically engineered Sake yeasts are  shown  in Figures 22a,b.  In contrast  to  the  fermentations of  a)  b)      56    Chardonnay must  in Section 3.1.3.1, Sake  fermentations were  less  robust, and  slower  (~600 hours  to  completion). The fermentation profiles shown in Figures 22a,b indicate substantial equivalence amongst  the parent and engineered strains as far as fermentation rate is concerned.  0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 0 100 200 300 400 500 600 T im e   (Hours) W ei g h t  lo ss  ( g ) K 7 K 7E C ‐   0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 0 100 200 300 400 500 600 T ime   (Hours) W ei g h t  lo ss  ( g ) K 9 K 9E C ‐   Figure 22. Fermentation profiles  (weight  loss) of parental and DUR1,2 engineered Sake strains  in Sake  wine. Sake wine was produced by inoculation (final OD600 = 0.1) of Sake yeast strains (a) K7 and K7 EC‐ and  (b)  K9  and  K9EC‐  into white  rice  (Kokako  Rose)  and  koji mash  (Vision  Brewing).  Fermentations were  incubated  to  completion  (~600  hours)  at  18°C.  Fermentations were  conducted  in  triplicate  and  data  averaged; error bars indicate one standard deviation.  3.1.3.3  Utilization of glucose and fructose and production of ethanol by K7EC‐ and K9EC‐ in Sake wine.  The effect of  the DUR1,2 cassette on glucose and  fructose utilization as well as ethanol production  in  a)  b)      57    strains K7EC‐ and K9EC‐ was investigated. Glucose, fructose and ethanol were quantified by LC analysis at  the end of  fermentation. Compared  to  their respective parental strains, K7EC‐ and K9EC‐ produced Sake  wine with substantially equivalent amounts of residual glucose, residual fructose and ethanol (Table 11).    Table 11. Utilization of glucose and fructose and production of ethanol  in Sake wine by parental yeast  strains  (K7  and  K9),  their  metabolically  engineered  counterparts  (K7EC‐  and  K9EC‐).  Sake  wine  was  produced by inoculation of Sake yeast strains, K7, K7EC‐, K9 and K9EC‐ in white rice (Kokako Rose) and koji  mash  (Vision Brewing).  Fermentations  (Figures  22a,b) were  incubated  to  completion  (~600 hours)  at  18°C. Glucose and  fructose  (g/L) and ethanol  (v/v) were quantified at  the end of  fermentation. Data  were analyzed for statistical significance (p≤0.05) using two factor ANOVA analysis.                                    3.1.4  Constitutive expression of DUR1,2 in Sake yeast strains K7EC‐ and K9EC‐ reduces EC in Chardonnay  wine by approximately 30%     In order to assess the effect of the DUR1,2 cassette on EC reduction  in Chardonnay wine, K7EC‐  and K9EC‐ were used to ferment Chardonnay must, and EC content was measured in the resultant wine.    Glucose    K7  K7EC‐  p* K9  K9EC‐  p*  Replicate 1  0.059  0.056  ‐‐  0.057  0.102  ‐‐  Replicate 2  0.066  0.056  ‐‐  0.114  0.103  ‐‐  Replicate 3  0.059  0.051  ‐‐  0.072  0.090  ‐‐  Residual glucose average (n=3)  0.06  0.05  ns 0.08  0.10  ns  STDEV  0.00  0.00  ‐‐  0.03  0.01  ‐‐        Fructose    K7  K7EC‐  p* K9  K9EC‐  p*  Replicate 1  0.736  0.733  ‐‐  0.333  0.242  ‐‐  Replicate 2  0.841  0.752  ‐‐  0.271  0.361  ‐‐  Replicate 3  0.374  0.665  ‐‐  0.338  0.303  ‐‐  Residual fructose average (n=3) 0.65  0.72  ns 0.31  0.30  ns  STDEV  0.25  0.05  ‐‐  0.04  0.06  ‐‐        Ethanol    K7  K7EC‐  p* K9  K9EC‐  p*  Replicate 1  12.23  12.49  ‐‐  12.71  13.22  ‐‐  Replicate 2  12.27  12.64  ‐‐  13.59  12.71  ‐‐  Replicate 3  11.95  11.93  ‐‐  13.05  12.17  ‐‐  Ethanol average (n=3)  12.15  12.35 ns 13.12  12.70  ns  STDEV  0.17  0.37  ‐‐  0.44  0.53  ‐‐  * si, ns: significant at p≤0.05, or non‐significant      58    Fermentation profiles of  the parental and metabolically engineered Sake yeasts are  shown  in Figures  21a,b.    In contrast to the usually high amounts of EC found  in Sake wine, during  laboratory scale wine  fermentations  the parental Sake  strains K7 and K9 produced  relatively  low amounts of EC, 64.87 and  56.16  ppm,  respectively  (Table  12);  the  engineered  strains  K7EC‐  and  K9EC‐  reduced  by  EC  in  the  Chardonnay wine by 29.88% and 0%, respectively.    Table 12. Reduction of EC by functionally enhanced yeast strains K7EC‐ and K9EC‐ during wine making. The  concentration of EC  (µg/L)  in Chardonnay wine produced by Sake yeast strains K7, K7EC‐, K9, and K9EC‐  was quantified by GC/MS. Strains were inoculated (final OD600 = 0.1) into unfiltered Calona Chardonnay  must and fermentations were incubated to completion (~350 hours) at 20°C. Fermentation profiles are  given in Figure 21.    Yeast strain  K7  K7EC‐  K9  K9EC‐  Replicate 1  69.05  46.3  56.13  55.18  Replicate 2  63.62  42.94  ‐‐  ‐‐  Replicate 3  61.93  47.21  ‐‐  ‐‐  Average (n=3)  64.87  45.48  56.16  55.18  STDEV  3.72  2.25  ‐‐  ‐‐  RSD (%)  5.73  4.94  ‐‐  ‐‐  Reduction (%)  ‐‐  29.88  ‐‐  ~0      3.1.5   Constitutive expression of DUR1,2 in Sake yeast strains K7EC‐ and K9EC‐ reduces EC in Sake wine  by approximately 68%     To assess the effect of the DUR1,2 cassette on EC reduction  in Sake wine, K7EC‐ and K9EC‐ were  used to ferment Sake rice mash (rice and koji), and EC content was measured in the resultant Sake wine.  Fermentation profiles of  the parental and metabolically engineered Sake yeasts are  shown  in Figures  22a,b.     During laboratory scale Sake wine fermentations the parental Sake strains K7 and K9 produced  significantly more EC (211.19 and 344.16 ppm, respectively ‐ Table 13) than during wine fermentations  (Table 12).  In Sake wine,  the engineered strains K7EC‐ and K9EC‐  reduced by EC by 67.54% and 68.33%,  respectively (Table 13), making K7EC‐ and K9EC‐ much more effective at EC reduction in Sake wine than in      59    Chardonnay wine. Given the substantial difference in EC production and reduction by identical strains, it  seems  imperative  that yeast strains be evaluated  in  their native environment  to accurately assess  the  efficacy of the DUR1,2 cassette.    Table 13. Reduction of EC by  functionally enhanced yeast strains K7EC‐ and K9EC‐ during Sake brewing.  The concentration of EC  (µg/L)  in Sake wine, produced by  inoculation  (final OD600 = 0.1) of Sake yeast  strains  K7,  K7EC‐,  K9,  and  K9EC‐  into  white  rice  (Kokako  Rose)  and  koji mash  (Vision  Brewing),  was  quantified  by  GC/MS.  Fermentations  were  incubated  to  completion  (~600  hours)  at  18°C  and  fermentation profiles are given in Figure 22.                  3.2  Constitutive expression of DUR3 in the Sake yeast strain K7 and the wine yeast strain 522     3.2.1  Sequence of pUCMD    A multicopy episomal plasmid containing the DUR3 ORF  inserted between the PGK1 promoter  and terminator signals flanked by TRP1 sequences was constructed; kanMX served as a selective marker  in  pUCMD  (Figure  14).  Single  strand  sequencing  revealed  this  plasmid  contained  the  desired  DNA  fragments  in  the  correct  order  and  orientation.  Furthermore,  in  silico  assembly  of  the DUR3  coding  region in pUCMD revealed that the DUR3 ORF was identical in amino acid sequence and length to that  published on SGD.    Aligning  the  sequence  data  of  the DUR3  cassette with  the  expected  sequence  revealed  nine  single nucleotide changes along  the  length of  the cassette  (Table 14), which are highlighted  in a DNA  sequence  alignment  between  of  S288C  and  pUCMD  (Figure  23).  A  detailed  description  of  the  DNA  sequences that comprise the DUR3 cassette is given in Table 15.     Yeast strain  K7  K7EC‐  K9  K9EC‐  Replicate 1  224.11  72.5  241.73  108.92  Replicate 2  198.7  79.28  265.01  105.96  Replicate 3  210.75  53.85  525.74  112.15  Average (n=3)  211.19  68.54  344.16  109.01  STDEV  12.71  13.17  157.68  3.10  RSD (%)  6.02  19.22  45.82  2.84  Reduction (%)  ‐‐  67.54  ‐‐  68.33      60    Table 14. Discrepancies between  the DUR3  cassette  in pUCMD  and published  sequences. Nucleotide  mismatches are designated as ‘XÆ Y’, meaning that the predicted nucleotide ‘X’ has been sequenced as  ‘Y’. The DNA alignment of S288C and pUCMD is given in Figure 23.    Nucleotide  position  Region of cassette  Description  928  PGK1 promoter  CÆT   999  PGK1 promoter  CÆT  1230  PGK1 promoter  CÆT  1491  PGK1 promoter  CÆT  1494  PGK1 promoter  GÆA  3524  DUR3 ORF  GÆC: Silent  3821  DUR3 ORF  TÆC: Silent  3970  DUR3 ORF  GÆA: Silent  4160  DUR3 ORF  AÆC: Silent    Table 15.  Detailed description of the DNA sequences that comprise the DUR3 cassette.    Nucleotide  position  Description  1‐6  5’ Apa1 restriction site  7‐291  5’ TRP1 sequence  292‐475  pHVXKD3 vector sequence from cloning strategy  476‐1976  PGK1 promoter  1977‐4184  DUR3 ORF  4185‐4467  PGK1 terminator  4468‐4535  pHVXKD3 vector sequence from cloning strategy  4536‐4933  kanMX promoter  4934‐5743  kanMX ORF  5744‐5992  kanMX terminator  5993‐6120  pHVXKD3 vector sequence from cloning strategy  6121‐6510  3’ TRP1 sequence  6511‐6516  3’ Apa1 restriction site                                  61    Alignment of DNA sequences: S288C and pUCMD    Upper line: S288C, from 1 to 6515  Lower line: pUCMD, from 1 to 6515    Data identity= 99%  901 TAGCATACAATTAAAACATGGCGGGCACGTATCATTGCCCTTATCTTGTGCAGTTAGACG ||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||| 901 TAGCATACAATTAAAACATGGCGGGCATGTATCATTGCCCTTATCTTGTGCAGTTAGACG 961 CGAATTTTTCGAAGAAGTACCTTCAAAGAATGGGGTCTCATCTTGTTTTGCAAGTACCAC |||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||| 961 CGAATTTTTCGAAGAAGTACCTTCAAAGAATGGGGTCTTATCTTGTTTTGCAAGTACCAC ------------------------------------------------------------ ---------------CONTINUATION OF PGK1 PROMOTER---------------- ------------------------------------------------------------ 1201 TCAAGACGCACAGATATTATAACATCTGCACAATAGGCATTTGCAAGAATTACTCGTGAG |||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||| 1201 TCAAGACGCACAGATATTATAACATCTGCATAATAGGCATTTGCAAGAATTACTCGTGAG ------------------------------------------------------------ ---------------CONTINUATION OF PGK1 PROMOTER---------------- ------------------------------------------------------------ 1441 CCGTCGCTCGTGATTTGTTTGCAAAAAGAACAAAACTGAAAAAACCCAGACACGCTCGAC |||||||||||||||||||||||||||||||||||||||||||||||||| || |||||| 1441 CCGTCGCTCGTGATTTGTTTGCAAAAAGAACAAAACTGAAAAAACCCAGATACACTCGAC 1501 TTCCTGTCTTCCTATTGATTGCAGCTTCCAATTTCGTCACACAACAAGGTCCTAGCGACG |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 1501 TTCCTGTCTTCCTATTGATTGCAGCTTCCAATTTCGTCACACAACAAGGTCCTAGCGACG ------------------------------------------------------------ ----CONTINUATION OF PGK1 PROMOTER AND START OF DUR3 ORF----- ------------------------------------------------------------ 3481 CTTTGCTATCACCAGCCATTTTTATTCCTATTTTAACGTATGTGTTTAAGCCACAAAATT ||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||| 3481 CTTTGCTATCACCAGCCATTTTTATTCCTATTTTAACGTATGTCTTTAAGCCACAAAATT ------------------------------------------------------------ ---------------CONTINUATION OF DUR3 ORF--------------------- ------------------------------------------------------------ 3781 TACAAAATGAATTAGACGAAGAACAAAGAGAACTAGCACGTGGTTTAAAAATTGCATACT |||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||| 3781 TACAAAATGAATTAGACGAAGAACAAAGAGAACTAGCACGCGGTTTAAAAATTGCATACT 3841 TCCTATGTGTTTTTTTCGCTTTGGCATTTTTGGTAGTTTGGCCCATGCCCATGTATGGTT |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 3841 TCCTATGTGTTTTTTTCGCTTTGGCATTTTTGGTAGTTTGGCCCATGCCCATGTATGGTT 3901 CCAAATATATCTTCAGTAAAAAATTCTTTACCGGTTGGGTTGTTGTGATGATCATCTGGC |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 3901 CCAAATATATCTTCAGTAAAAAATTCTTTACCGGTTGGGTTGTTGTGATGATCATCTGGC 3961 TTTTTTTCAGTGCGTTTGCCGTTTGTATTTATCCACTCTGGGAAGGTAGGCATGGTATAT     62    ||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||| 3961 TTTTTTTCAGTGCATTTGCCGTTTGTATTTATCCACTCTGGGAAGGTAGGCATGGTATAT 4021 ATACCACTTTGCGAGGCCTTTACTGGGATCTATCTGGTCAAACTTATAAATTAAGGGAAT |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 4021 ATACCACTTTGCGAGGCCTTTACTGGGATCTATCTGGTCAAACTTATAAATTAAGGGAAT 4081 GGCAAAATTCGAACCCACAAGATCTGCATGTAGTAACAAGCCAAATTAGTGCGAGAGCAC |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 4081 GGCAAAATTCGAACCCACAAGATCTGCATGTAGTAACAAGCCAAATTAGTGCGAGAGCAC 4141 ATAGACAATCATCACATTTCGGACAAGTTGATGAAATAATTTAGCTCGAGGATTGAATTG |||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||| 4141 ATAGACAATCATCACATTTCGGCCAAGTTGATGAAATAATTTAGCTCGAGGATTGAATTG 4201 AATTGAAATCGATAGATCAATTTTTTTCTTTTCTCTTTCCCCATCCTTTACGCTAAAATA |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 4201 AATTGAAATCGATAGATCAATTTTTTTCTTTTCTCTTTCCCCATCCTTTACGCTAAAATA Figure 23.  DNA sequence alignment of S288C and pUCMD revealed nine discrepancies along the length  of DUR3 cassette. The sequence obtained from S288C  is shown  in the upper  line and the sequence of  the DUR3 cassette in pUCMD is shown in the lower line. Only a partial sequence of the DUR3 cassette is  displayed.  The  PGK1  promoter  is  shown  in  green,  the  DUR3  ORF  is  shown  in  black,  and  the  PGK1  terminator is shown in blue. Nucleotide mismatches are highlighted in bold and red, and are underlined.  In  the  DUR3 ORF, mismatches  are  oriented within  their  respective  codons  by  underlining;  the  stop  codon is italicized.       In  silico  analysis  of  the  integrated DUR3  cassette  revealed  that  two  new ORFs were  created  during  construction  of  the DUR3  cassette;  these ORFs were  composed  of  a mixture  of  S.  cerevisiae  sequence and pHVXK vector sequence (Figure 24). Novel ORF1 (324 bp) is located at nucleotide position  7‐330 while novel ORF2 (621 bp) is located at position 463‐1083.  Figure  24.  A  schematic  representation  of  new  ORFs  of  more  than  100  codons  generated  during  construction of the DUR3 cassette. Two new ORFs, composed of a mixture of S. cerevisiae sequence and  pHVXK vector sequence, were created.  Novel ORF2  (621 bp)  Novel ORF1  (324 bp)  5’ TRP1  PGK1p  kanMX  3’ TRP1PGK1tDUR3 kanMXp  kanMXt 5’ Apa1 site  3’ Apa1 site     63    3.2.2  Integration of the linear DUR3 cassette into the genomes of yeast strains K7, K7EC‐, K9, and K9EC‐  To constitutively express DUR3, the linear DUR3 cassette (Figure 25) was transformed into Sake  yeast strains K7 and K7EC‐, and the wine strains 522 and 522EC‐. Following positive selection on a G418  medium and sub‐culturing, the eight strains listed in Table 16 were obtained.    Table 16. Recombinant yeast strains created by  integration of  the DUR3 cassette  into  the TRP1  locus.  The  designation  ‘EC‐‘  corresponds  to  integration  of  the  DUR1,2  cassette  while  ‘D3’  corresponds  to  integration of the DUR3 cassette. ‘EC‐D3’ designates integration of both cassettes.    Genetic modification  Parent strain  K7  522  Wild type  K7  522  DUR1,2  K7EC‐  522EC‐  DUR3  K7D3  522D3  DUR1,2/DUR3  K7EC‐D3  522EC‐D3    3.2.3  Genetic characterization of K7D3, K7EC‐D3, 522D3, and 522EC‐D3    3.2.3.1  Correct integration of the DUR3 cassette into the genomes of K7D3, K7EC‐D3, 522D3, and 522EC‐D3.  A Southern blot, using EcoR1 digested genomic DNA and PCR created DUR3 and TRP1 probes (Figure 26),  was  performed  on  each  of  the  recombinant  yeast  strains  to  assess  proper  integration  of  the DUR3  cassette into the genome. The DUR3 and TRP1 probes are schematically represented in Figures 27a,b.      In each of the recombinant yeast strains, two signals corresponding to 2.4 kb and 4.7 kb, were  detected when probed for DUR3 (Figure 26). The 2.4 kb fragment matches the expected fragment size  for  the native DUR3  locus while  the 4.7 kb  fragment matches  the expected  size  for  the  recombinant  TRP1‐PGK1p‐DUR3‐PGK1t‐kanMX‐TRP1  locus.  In  strains  that  did  not  carry  the  recombinant  DUR3  cassette, only the 2.4 kb signal was detected.   Figure 25. Schematic representation of the linear DUR3 cassette.    TRP1  PGKp  kanMX  TRP1 PGKt DUR3  Apa1 site  Apa1 site      64      When probed for TRP1, three signals, corresponding to 1.5 kb, 3.0 kb and 4.7 kb, were detected  in  each  of  the  strains  containing  the  recombinant  DUR3  cassette  (Figure  26).  The  1.5  kb  fragment  matches  the  expected  fragment  size  for  a  non‐disrupted  TRP1  locus,  while  the  3.0  kb  and  4.7  kb  fragments are  in accordance with  the presence of the recombinant DUR3 cassette  integrated  into the  TRP1 locus.     Unexpectedly, when any of the K7 derived strains, except for K7EC‐D3, were probed for DUR3, a  novel signal was detected at ~3.5 kb (Figure 26). This fragment corresponds to the disappearance of the  2.4  kb  band  that  represents  the  native DUR3  locus.  This  result was  confirmed  by  sequencing  to  be  caused by a restriction fragment length polymorphism (RFLP) between 522 and K7. The DUR3 locus in K7  is mutated such that it no longer contains the EcoR1 site within the coding region (Figure 28) and, as a  Figure 26. Integration of the DUR3 cassette into the TRP1 locus of 522D3, 522EC‐D3, K7D3, and K7EC‐D3 was  confirmed  by  Southern  blot  analysis  using  DUR3  and  TRP1  probes.  Approximately  1  µg  of  EcoR1  digested genomic DNA from 522 (lane #1), 522D3 (lane #2), 522EC‐D3 (lane #3), K7 (lane #4), K7EC‐ (lane  #5), K7D3 (lane #6), and K7EC‐D3 (lane #7) was probed with either DUR3 or TRP1. In lane #5 less DNA was  accidentally loaded thus accounting for the faint band pattern. All faint bands that did not correspond  to a size in the schematic representation were deemed to be non‐specific binding.    4680 bp DUR3 probe TRP1 probe 1452 bp 3088 bp 4680 bp 2444 bp ~ 3500 bp 1 2 3 4 5 6 7 1 2 3 4 5 6 7     65    EcoR1 a)  DUR3 PGK1p trp1 trp1 PGK1t DUR3 EcoR1 2444 bp EcoR1 EcoR1 4680 bp   EcoR1 3088 bp EcoR1 DUR3 PGK1p trp1 trp1 PGK1t TRP1 EcoR1 1452 bp EcoR1 EcoR1 4680 bp   b)  result, the native DUR3  locus  in K7 produced a  larger signal (~3500 bp). By  integrating a copy of DUR3  derived from 522, which contains the internal EcoR1 site, a recombinant locus was created which gives  rise to the same band pattern seen  in 522. Although this RFLP should have been observed  in K7EC‐D3  it  was not and therefore warrants further investigation.                  Figure  27.  Schematic  representation  of  the  signals  expected  during  Southern  blot  analysis  of  recombinant  yeasts  containing  the  recombinant  DUR3  cassette  integrated  into  the  TRP1  locus.  Expected  signals  during  Southern  blotting with  the  DUR3  probe  (drawn  in  green)  (a).  Expected  signals during Southern blotting with the TRP1 probe (drawn in red) (b).        66    Alignment of DNA sequences: S288C, 522, K7 (1‐240 bp DUR3)  S288C ATGGGAGAATTTAAACCTCCGCTACCTCAAGGCGCTGGGT 40 522 ......................cAtCcaggGGCGCTGGGa 18 K7 ............cAACaaaCcCcctactcAGGCGCTGGGT 28 S288C ACGCTATTGTATTGGGCCTAGGGGCCGTATTTGCAGGAAT 80 522 .CGCTATTGTATTGGGCCTAGGGGCCGTATTTGCAGGAAT 57 K7 .CGCTATTGTATTGGGCCTAGGGGCCGTATTTGCAGGAAT 67 S288C GATGGTTTTGACCACTTATTTACTGAAACGTTATCAAAAG 120 522 GATGGTTTTGACCACTTATTTACTGAAACGTTATCAAAAG 97 K7 GATGGTTTTGACCACTTATTTACTGAAACGTTATCAAAAG 107 S288C GAAATCATCACAGCAGAAGAATTCACCACCGCCGGTAGAT 160 522 GAAATCATCACAGCAGAAGAATTCACCACCGCCGGTAGAT 137 K7 GAAATCATCACAGCAGAAGAATTtACCACCGCCGGcAGAT 147 S288C CTGTAAAAACCGGCTTAGTGGCTGCAGCCGTGGTTTCTAG 200 522 CTGTAAAAACCGGCTTAGTGGCTGCAGCCGTGGTTTCTAG 177 K7 CTGTAAAAACCGGCTTAGTGGCTGCcGCCGTGGTTTCTAG 187 S288C TTGGATCTGGTGTTCTACATTGTTAACGTCGTCAACAAAG 240 522 TTGGATCTGGTGTTCTACATTGTTAACGTCGTCAACAAAG 217 K7 TTGGATCTGGTGTTCTACATTGTTAACGTCGTCAACAAAG 227   Figure 28. Alignment of the DNA sequences of S288C, 522, and K7 confirmed the presence of a mutant  EcoR1 site in the DUR3 coding region of K7. Only a partial sequence of DUR3 is displayed. The EcoR1 site  in question is displayed in red; other mismatches revealed by sequencing are highlighted in blue.      3.2.3.2   Confirmation of constitutive expression of DUR3  in K7D3 and K7EC‐D3 by northern blotting.  In  order  to  assess  the  constitutive  expression  of  DUR3  from  the  DUR3  cassette  in  K7D3  and  K7EC‐D3,  a  northern blot was performed using PCR generated probes for HHF1 (Histone H4) and DUR3 on freshly  harvested total RNA from K7, K7EC‐, K7D3, and K7EC‐D3. Identical signals matching the predicted transcript  size for HHF1 (312 bp) were observed in all of the strains tested (Figure 29), indicating an abundance of  un‐degraded mRNA  in the samples. Signals, which matched the predicted transcript size  (2208 bp) for  DUR3, were observed  in  strains K7D3  and K7EC‐D3 when probed  for DUR3  (Figure 29),  thus  confirming  constitutive  expression  of  DUR3  in  non‐inducing  conditions  as  a  result  of  integration  of  the  DUR3  cassette. No DUR3  signal was  detected  in  K7  or  K7EC‐  confirming  that DUR3 mRNA  is  absent  during  fermentation due to repression by NCR.        67                                3.2.3.3   Quantification of constitutive DUR3 expression  in K7D3 and K7EC‐D3 by qRT‐PCR. Expression of  DUR3 and DUR1,2 was verified and quantified by qRT‐PCR analysis of cDNA reverse transcribed from the  total  RNA  of  K7,  K7EC‐,  K7D3,  and  K7EC‐D3.   Analyses were  performed  on  the  same  RNA  samples  as  in  Section 3.2.3.2.    The PGK1 promoter and terminator signals resulted in high level expression, under non‐inducing  (NCR) conditions, of both the DUR1,2 and DUR3 genes  in the  integrated cassettes (Figure 30); DUR1,2  was upregulated 11.8‐fold  in K7EC‐ while DUR3 was upregulated 22.1‐fold  in K7D3. High  level expression  of both DUR1,2 and DUR3  (17.3‐fold and 11.5‐fold,  respectively) was  sustained  in K7EC‐D3, when both  DUR1,2 and DUR3 were integrated into the genome.      In strains in which only one cassette (DUR1,2 or DUR3) was integrated, constitutive expression  of  that  gene  induced  expression  of  the  other  (Figure  30).  Constitutive  expression  of DUR1,2  in  K7EC‐  Figure 29. Constitutive expression of DUR3  (2208 bp) was confirmed by northern blot analysis of K7,  K7EC‐, K7D3, K7EC‐D3. Total RNA (30 µg) was harvested from 24 hour fermentations of Calona Chardonnay  must inoculated to a final OD600 = 0.1, separated on a 1% agarose‐formaldehyde gel and probed for both  the loading control HHF1 and DUR3.    K7   K7 EC ‐   K7 D 3   K7 EC ‐D 3   K7   K7 EC ‐   K7 D 3   K7 EC ‐D 3   DUR3HHF1 312bp  2208bp      68    induced  expression  of DUR3  by  3.78‐fold.  Likewise,  constitutive  expression  of DUR3  in  K7D3  induced  expression of DUR1,2 by 3.61‐fold.   1.00 9.26 3.61 11.82 1.00 3.78 14.19 7.10 0 2 4 6 8 10 12 14 16 18 K7 K7EC- K7D3 K7EC-D3 Strain R el at iv e qu an tif ic at io n (fo ld ) DUR1,2 DUR3 R el at iv e qu an tif ic at io n (fo ld )   Figure 30. Analyses of gene expression (qRT‐PCR) of K7, K7EC‐, K7D3, and K7EC‐D3 confirmed functionality of  the DUR3 cassette and constitutive expression of DUR1,2 and DUR3  in non‐inducing  (NCR) conditions.  Total RNA was extracted from cells harvested after 24 hour fermentation (20°C) in filter sterilized Calona  Chardonnay must  that  was  inoculated  to  a  final  OD600  =  0.1.  Total  RNA  was  subsequently  reverse  transcribed, and the resultant cDNA was amplified in the presence of SYBR green dye. DUR1,2 and DUR3  gene expression was standardized to ACT1 expression and data for K7EC‐, K7D3, K7EC‐D3 was calibrated to  the parental  strain K7. Fermentations were  conducted  in  triplicate and  the data averaged; error bars  represent 95% confidence intervals.      3.2.3.4  Effect of the integrated DUR3 cassette on the transcriptome of K7D3. Total RNA from yeast cells  isolated from 24 hour fermentations in Chardonnay must was used analyze the impact of the integrated  DUR3 cassette on global gene expression in K7D3. Reported changes in gene expression were cut off at a  minimum 4‐fold change  in expression  (SLRAvg ≥ 2 or SLRAvg ≤  ‐2)  to ensure elimination of experimental  noise  inherent  in microarray  analysis;  this  cut‐off  is  supported by  the previously published  statistical  examination of a wine yeast’s transcriptome during fermentation (Marks, et al. 2008).         69    Integration of the DUR3 cassette into the TRP1 locus of K7D3 had a minimal effect on the yeast’s  transcriptome. DUR3 was upregulated by 36.95‐fold  in K7D3; TRP1 was downregulated by 1.25‐fold but  was  not  included  in  Table  17  as  it  fell  below  the  4‐fold  cut  off.  Besides  DUR3,  three  genes  were  upregulated  greater  than  4‐fold  in  K7D3  (Table  17);  one  gene  (HAC1)  was  common  to  both  of  the  engineered strains K7EC‐ and K7D3 (Tables 10 and 17). Four genes were downregulated more than 4‐fold  in K7D3. Despite falling below the 4‐fold cut off, one gene (RUD3) was included in Table 17 because it was  common to the list of genes downregulated in both K7EC‐ and K7D3 (Tables 10 and 17) and because it is  very close to the 4‐fold cut off. No metabolic pathways were affected by the presence of the integrated  DUR3  cassette;  however,  integration  of  the  DUR3  cassette  upregulated  one  gene  (FIG1)  and  downregulated  one  gene  (TID3)  involved  in  meiosis/sporulation  indicating  a  possible  effect  on  meiosis/sporulation efficiency (Table 17).      Table 17. Effect of the integrated DUR3 cassette in the genome of K7 on global gene expression patterns  in S. cerevisiae K7D3 (≥ 4‐fold change). Reported changes are relative to the parental strain K7. Total RNA  from K7 and K7D3, harvested at 24 hours  into  fermentation of  filter sterilized Chardonnay must, were  used  for hybridization  to microarrays. Fermentations were conducted  in duplicate and  the data were  averaged (p≤0.005).     Genes expressed at higher levels in K7D3 Fold Change Gene Symbol Biological Process 36.95 DUR3 Urea permease 5.67 HAC1 bZIP (basic-leucine zipper) protein involved in unfolded protein response 4.93 BRN1 Protein required for chromosome condensation 4.24 FIG1 Integral membrane protein required for efficient mating and low affinity Ca2+ transport Genes expressed at lower levels in K7D3 Fold Change Gene Symbol Biological Process -8.18 TID3 Meiotic protein required for synapsis and meiotic recombination; interaction partner with DMC1p -7.31 SNT309 Splicing factor protein -5.98 TOA1 Transcription factor IIA, large chain -3.98 RUD3 Protein involved in organization of Golgi   3.2.4  Recombinant strains K7D3 and K7EC‐D3 exhibit highly enhanced urea uptake ability in conditions of  strong NCR    Radiolabelled 14C‐urea uptake assays were performed to confirm constitutive urea uptake as a  result of integration of the DUR3 cassette.      70    The parental Sake  strain K7, and  the  recombinant  yeast K7EC‐  containing  the DUR1,2  cassette  integrated  into the URA3  locus, failed to  incorporate any significant amount of radiolabelled urea  in a  minimal medium containing 1% (w/v) ammonium sulphate (Figure 31). This  is  likely due to the NCR of  the DUR genes in the presence of the preferred nitrogen source (ammonium sulphate). In contrast, K7D3  and K7EC‐D3 were highly  efficient  at urea uptake  (Figure 31),  confirming  that  integration of  the DUR3  cassette results  in the production of a functional protein that  localizes to the yeast plasma membrane,  and that control of DUR3 by the PGK1 promoter and terminator signals is capable of overcoming native  repression by NCR.    A  significant difference  in  the urea uptake  rates of K7D3 and K7EC‐D3 was observed  (Figure 31),  despite  integration of  identical DUR3 cassettes  in both strains. The observed difference  in urea uptake  between K7D3 and K7EC‐D3 can be explained by the need for urea degradation (DUR1,2p) after its import  by  the  cell. While  DUR1,2 must  be  induced  and  then  synthesized  in  K7D3,  DUR1,2p  is  constitutively  expressed in K7EC‐D3 leading to rapid degradation of urea to which might mask an increased uptake; the  radiolabel would be quickly lost as 14CO2.        71      Figure  31.   Uptake  of  14C‐urea  by  K7,  K7EC‐,  K7D3  and  K7EC‐D3  under  conditions  of NCR.    Strains were  cultured (final OD600 = 1) in a 1% (w/v) ammonium sulphate minimal medium prior to exposure to 0.27  mM  14C‐urea  (6.8 mCi/mmol). Assays were  conducted  in  triplicate  and  the data  averaged; error bars  represent one standard deviation.      When compared to either K7D3 or K7EC‐D3, both K7 and K7EC‐ exhibited relatively poor urea uptake  under the conditions tested  (Figure 31).  In order  to differentiate between K7 and K7EC‐, Figure 31 was  modified by significantly reducing the scale of the Y‐axis; the maximum value of the Y‐axis was reduced  from 7.0 nmole (Figure 31) to 0.1 nmole (Figure 32). K7 was capable of accumulating significantly more  urea  than  K7EC‐  (Figure  32),  probably  due  to  the  constitutive  expression  of DUR1,2  in  K7EC‐;  14C‐urea  might be actively degraded as  it  is  imported  into  the cell  thus giving  the  illusion  that  transport  is  less  efficient. This phenomenon was also observed when  the abilities of K7D3 and K7EC‐D3  to  transport urea  into the cell were compared (Figure 31).       72        Figure 32. Uptake of 14C‐urea by K7 and K7EC‐ under conditions of NCR. Strains were cultured (final OD600  = 1)  in a 1%  (w/v) ammonium  sulphate minimal medium prior  to exposure  to 0.27 mM  14C urea  (6.8  mCi/mmol).  Assays  were  conducted  in  triplicate  and  the  data  averaged;  error  bars  represent  one  standard deviation.      3.2.5  Recombinant strains K7D3 and K7EC‐D3 exhibit highly enhanced urea uptake ability in conditions of  NCR de‐repression    Radiolabelled  14C‐urea  uptake  assays  were  performed  in  order  to  assess  the  ability  of  engineered  strains  containing  the  integrated  DUR3  cassette  to  uptake  urea  under  non‐repressive  conditions.       73    Both  the  parental  Sake  strain  K7,  and  the  recombinant  yeast  K7EC‐  containing  the  DUR1,2  cassette  integrated  into  the URA3  locus, did not  incorporate  any  significant  amount of  14C‐urea  in  a  minimal medium containing 1% (w/v) L‐proline (Figure 33). This result, which parallels the inability of K7  and K7EC‐ to incorporate 14C‐urea in an ammonium sulphate minimal medium (Figure 31), is likely due to  the  slow  induction and membrane  trafficking of  functional DUR3p.  In  contrast, K7D3 and K7EC‐D3 were  highly efficient at urea uptake (Figure 33), confirming that integration of the DUR3 cassette results in the  production  of  a  functional  protein,  and  that  transcription  of  DUR3  from  the  PGK1  promoter  and  terminator signals is strong regardless of NCR state.        Figure  33.  Uptake  of  14C‐urea  by  K7,  K7EC‐,  K7D3  and  K7EC‐D3  under  conditions  of  NCR  de‐repression. Strains were cultured (final OD600 = 1) in a 1% (w/v) L‐proline minimal medium prior to exposure to 0.27  mM  14C urea  (6.8 mCi/mmol). Assays were  conducted  in  triplicate and  the data averaged; error bars  represent one standard deviation.      74    When compared to either K7D3 or K7EC‐D3, both K7 and K7EC‐ exhibited relatively poor urea uptake  under conditions of NCR de‐repression  (Figure 33), K7 was capable of accumulating significantly more  urea than K7EC‐ (Figure 34). To differentiate between K7 and K7EC‐, Figure 33 was modified by reducing  the scale of the Y‐axis; the maximum value of the Y‐axis was reduced from 7.0 nmole (Figure 33) to 0.1  nmole  (Figure 34). K7 was capable of accumulating significantly more urea  than K7EC‐  (Figure 34);  14C‐ urea might be actively degraded as  it  is  imported  into K7EC‐ where DUR1,2  is constitutively expressed,  thus masking the strain’s true ability to import urea.     Figure  34.  Uptake  of  14C‐urea  by  K7  and  K7EC‐  under  conditions  of  NCR  de‐repression. Strains were  cultured (final OD600 = 1) in a 1% (w/v) L‐proline minimal medium prior to exposure to 0.27 mM  14C urea  (6.8 mCi/mmol). Assays were conducted  in  triplicate and  the data averaged; error bars represent one  standard deviation.        75    3.2.6  Phenotypic characterization of K7D3, K7EC‐D3, 522D3, and 522EC‐D3    3.2.6.1   Fermentation rate of K7D3, K7EC‐D3, 522D3, and 522EC‐D3  in Chardonnay wine. The  fermentation  profiles of  the parental  and metabolically  engineered  strains  are  shown  in  Figures 35a,b.  The  robust  fermentations were completed within 300 hours, and the fermentation rates of all but one engineered  strain closely matched those of their respective parental strains, thus indicating substantial equivalence.   The engineered  strain K7EC‐D3  completed  the  fermentation by approximately 200‐250 hours while  the  strains K7EC‐, and K7D3  required 300 hours  to  finish;  the parental  strain K7 also  required 300 hours  to  complete the fermentation.        76    0.00 5.00 10.00 15.00 20.00 25.00 0 50 100 150 200 250 300 350 T im e   (Hours) W ei g h t  lo ss  ( g ) K 7 K 7  E C ‐ K 7  D3 K 7  E C ‐D3   0.00 5.00 10.00 15.00 20.00 25.00 0 50 100 150 200 250 300 350 T im e   (Hours) W ei g h t  lo ss  ( g ) 522 522  E C ‐ 522  D3 522  E C ‐D3   Figure 35. Fermentation profiles (weight loss) of (a) Sake yeast strains K7, K7EC‐, K7D3, and K7EC‐D3 and (b)  wine yeast strains 522, 522EC‐, 522D3, and 522EC‐D3 in Chardonnay wine. Chardonnay wine was produced  from unfiltered Calona Chardonnay must inoculated to a final OD600 = 0.1 and incubated to completion  (~300 hours) at 20°C. Fermentations were conducted  in  triplicate and data were averaged; error bars  indicate one standard deviation.      b)  a)      77    3.2.6.2   Ethanol production by K7D3, K7EC‐D3, 522D3, and 522EC‐D3  in Chardonnay wine. The effect of the  DUR3 cassette on ethanol production in Chardonnay wine by strains K7D3, K7EC‐D3, 522D3, and 522EC‐D3 was  investigated by LC analysis at  the end of  fermentation. Compared  to  their  respective parental strains,  K7D3,  K7EC‐D3,  522D3,  and  522EC‐D3  produced  Chardonnay  wine  with  substantially  equivalent  ethanol  content (Table 18).    Table 18. Ethanol produced by Sake yeast  strains  (K7, K7EC‐, K7D3, and K7EC‐D3) and wine yeast  strains  (522, 522EC‐, 522D3, and 522EC‐D3) in Chardonnay wine. Ethanol content (%v/v) was measured by LC at the  end of fermentation. Fermentation profiles are given in Figures 35a,b. Data were analyzed for statistical  significance (p≤0.05) using two factor ANOVA analysis.        3.2.6.3  Fermentation rate of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 in Sake wine. The fermentation profiles of  the  parental  and  metabolically  engineered  strains  are  shown  in  Figures  36a,b.  In  all  cases,  the  fermentations proceeded  at  a  slower  rate  than  those  in  grape must;  fermentations were  completed  within 450 hours compared to 300 hours in Chardonnay grape must (Figure 35). The fermentation rates  of  engineered  strains  closely  matched  those  of  the  respective  parental  strains,  thus  indicating  substantial equivalence in fermentation rate.      K7  K7EC‐  K7D3  K7EC‐D3  p*  Replicate 1  12.89  12.84  12.89  12.91  ‐‐  Replicate 2  12.90  12.75  13.01  12.95  ‐‐  Replicate 3  12.96  12.97  12.97  12.89  ‐‐  Ethanol average (n=3)  12.92  12.85  12.96  12.92  ns  STDEV  0.04  0.11  0.06  0.03  ‐‐                522  522EC‐  522D3  522EC‐D3  p*  Replicate 1  13.65  13.71  13.74  13.54  ‐‐  Replicate 2  13.60  13.65  13.71  13.62  ‐‐  Replicate 3  13.71  13.66  13.55  13.58  ‐‐  Ethanol average (n=3)  13.65  13.67  13.67  13.58  ns  STDEV  0.06  0.03  0.10  0.04  ‐‐  * si, ns: significant at p≤0.05, or non‐significant      78    0.00 2.00 4.00 6.00 8.00 10.00 12.00 0 100 200 300 400 500 Hours W ei g h t  lo ss  ( g ) K 7 K 7  E C ‐ K 7  D3 K 7  E C ‐D3 0.00 2.00 4.00 6.00 8.00 10.00 12.00 0 100 200 300 400 500 Hours W ei g h t  lo ss  ( g ) 522 522 E C ‐ 522 D3 522 E C ‐D3   Figure 36. Fermentation profiles (weight loss) of (a) Sake yeast strains K7, K7EC‐, K7D3, and K7EC‐D3 and (b)  wine  yeast  strains  522,  522EC‐,  522D3,  and  522EC‐D3  in  Sake. Sake wine was  produced  from white  rice  (Kokako  Rose)  and  koji  mash  (Vision  Brewing)  inoculated  to  a  final  OD600  =  0.1  and  incubated  to  completion (~450 hours) at 18°C. Fermentations were conducted  in triplicate and data were averaged;  error bars indicate one standard deviation.    a)  b)      79    3.2.6.4   Ethanol production by K7D3, K7EC‐D3, 522D3, and 522EC‐D3  in Sake wine. The effect of  the DUR3  cassette on ethanol production in Sake wine by strains K7D3, K7EC‐D3, 522D3, and 522EC‐D3 was investigated  by LC analysis at  the end of  fermentation. Compared  to  their respective parental strains, K7D3, K7EC‐D3,  522D3, and 522EC‐D3 produced Sake wine with substantially equivalent ethanol content (Table 19).     Table 19. Ethanol produced by Sake yeast  strains  (K7, K7EC‐, K7D3, and K7EC‐D3) and wine yeast  strains  (522, 522EC‐, 522D3, and 522EC‐D3) in Sake wine. Ethanol content (% v/v) was measured by LC at the end of  fermentation.  Fermentation  profiles  are  given  in  Figures  36a,b.  Data  were  analyzed  for  statistical  significance (p≤0.05) using two factor ANOVA analysis.      K7  K7EC‐  K7D3  K7EC‐D3  p*  Replicate 1  9.33  9.22  8.90  8.63  ‐‐  Replicate 2  9.20  8.14  9.58  8.96  ‐‐  Replicate 3  10.01  8.14  9.64  9.01  ‐‐  Ethanol average (n=3)  9.51  8.50  9.37  8.87  ns  STDEV  0.44  0.62  0.41  0.21  ‐‐                522  522EC‐  522D3  522EC‐D3  p*  Replicate 1  8.78  7.46  8.77  8.99  ‐‐  Replicate 2  8.80  7.81  8.99  8.86  ‐‐  Replicate 3  8.61  9.22  8.98  8.37  ‐‐  Ethanol average (n=3)  8.73  8.16  8.91  8.74  ns  STDEV  0.10  0.93  0.12  0.33  ‐‐  * si, ns: significant at p≤0.05, or non‐significant      3.2.7  Constitutive expression of DUR3 in yeast strains K7D3 and 522D3 reduces EC in Chardonnay wine  by 24.97% and 81.38%, respectively.    In order to assess the reduction of EC by functionally enhanced yeast strains, Chardonnay wine  was made with the parental strains K7 and 522 and the recombinant strains K7EC‐, K7D3, K7EC‐D3, 522EC‐,  522D3,    and  522EC‐D3  and  the  EC  content  quantified  by  GC/MS  at  the  end  of  fermentation.  The  fermentation profiles are given in Figures 35a,b.  During wine making, K7D3 and 522D3  reduced EC as efficiently as K7EC‐ and 522EC‐,  respectively  (Table 20).  In Chardonnay wine,  K7D3  reduced  EC by 12.64% while K7EC‐  reduced  EC by  6.85%; 522D3      80    reduced EC by 82.96% while 522EC‐  reduced EC by 81.48%. Constitutive  co‐expression of DUR1,2 and  DUR3 did not result in synergistic EC reduction in K7EC‐D3 or 522EC‐D3 (Table 20).    Table 20. Reduction of EC by functionally enhanced yeast strains during wine making. The concentration  of EC  (µg/L)  in Chardonnay wine produced by Sake yeast  strains K7, K7EC‐, K7D3, and K7EC‐D3 and wine  yeast strains 522, 522EC‐, 522D3, and 522EC‐D3 from unfiltered Calona Chardonnay must was quantified by  GC/MS. Triplicate fermentations were  incubated to completion (~300 hours) at 20°C and fermentation  profiles are given in Figures 35a,b.    Yeast strain  K7  K7EC‐  K7D3  K7EC‐D3  Replicate 1  34.19 34.25 29.39 29.74 Replicate 2  40.31 30.98 29.98 31.29 Replicate 3  28.69 30.89 30.78 30.06 Average (n=3)  34.40 32.04 30.05 30.36 STDEV  5.81 1.91 0.70 0.82 RSD (%)  16.89 5.96 2.33 2.70 Reduction (%)  -- 6.85 12.64 11.73           Yeast strain  522  522EC‐  522D3  522EC‐D3  Replicate 1  199.3 31.32 28.76 36.63 Replicate 2  169.95 37.32 29.89 38.61 Replicate 3  177.61 32.66 34.56 31.3 Average (n=3)  182.29 33.77 31.07 35.51 STDEV  15.22 3.15 3.07 3.78 RSD (%)  8.34 9.33 9.88 10.64 Reduction (%)  -- 81.48 82.96 80.52     3.2.8   Constitutive  expression of DUR3  in  yeast  strains K7D3  and 522D3  reduces  EC  in  Sake wine by  18.40% and 10.45%, respectively.    To assay the EC reduction of functionally enhanced yeast strains during Sake making, Sake wine  was brewed with the parental strains K7 and 522 and the recombinant strains K7EC‐, K7D3, K7EC‐D3, 522EC‐,  522D3,  and  522EC‐D3  and  the  EC  content  quantified  by  GC/MS  at  the  end  of  fermentation.    The  fermentation profiles are given in Figures 36a,b.    As observed before  (Section 3.1.5), engineered  Sake  yeast  strains  reduce EC more effectively  when assessed by brewing  Sake;  the  Sake  yeasts K7EC‐ and K7EC‐D3 both  reduced EC  content by ~84%      81    (Table 21). During the wine making trial described in Section 3.2.7, K7EC‐ and K7EC‐D3 both reduced EC by  only ~15% (Table 20).    In contrast to the results described in Section 3.2.7 for Chardonnay wine, constitutive expression  of DUR3 did not reduce EC as effectively as constitutive expression of DUR1,2 (Tables 20 and 21). In Sake  wine, K7D3 reduced EC by 14.97% while K7EC‐ reduced EC by 87.07%; 522D3 reduced EC by 11.49% while  522EC‐  reduced  EC  by  84.30%  (Table  21).  Constitutive  co‐expression  of  DUR1,2  and  DUR3  had  no  synergistic  effect  on  EC  reduction  during  Sake  production  (Table  21);  there  was  no  appreciable  difference  in  EC  reduction  between  the  DUR1,2  expressing  strains  and  those  which  constitutively  expressed both DUR1,2 and DUR3.    Table 21. Reduction of EC by functionally enhanced yeast strains during Sake making. The concentration  of  EC  (µg/L)  in  Sake wine produced by  Sake  yeast  strains  K7,  K7EC‐,  K7D3,  and  K7EC‐D3  and wine  yeast  strains 522, 522EC‐, 522D3, and 522EC‐D3 from white rice (Kokako Rose) and koji mash (Vision Brewing) was  quantified by GC/MS. Triplicate fermentations were  incubated to completion (~450 hours) at 18°C and  the fermentation profiles are given in Figures 36a,b.    Yeast strain  K7  K7EC‐  K7D3  K7EC‐D3  Replicate 1  109.83 13.37 93.42 15.72 Replicate 2  80.68 12.3 79.86 18.6 Replicate 3  108.26 12.96 80.77 17.79 Average (n=3)  99.59 12.88 84.68 17.37 STDEV  16.40 0.54 7.58 1.49 RSD (%)  16.47 4.19 8.95 8.58 Reduction (%)  -- 87.07 14.97 82.56           Yeast strain  522  522EC‐  522D3  522EC‐D3  Replicate 1  77.05 11.6 85.85 16.12 Replicate 2  104.01 12.17 82.17 13.07 Replicate 3  85.11 18.03 67.57 13.51 Average (n=3)  88.72 13.93 78.53 14.23 STDEV  13.84 3.56 9.67 1.65 RSD (%)  15.60 25.56 12.31 11.60 Reduction (%)  -- 84.30 11.49 83.96           82    4  DISCUSSION    4.1  Constitutive expression of the DUR1,2 cassette reduces EC production in wine and Sake    A common theme throughout the history of mankind has been the lack of effective solutions to  problems until  the  advent of  certain  technologies.  Such has been  the  case  for  the problem of  EC  in  fermented  foods and beverages, more  specifically EC  in grape and Sake wine. Since  its discovery and  subsequent characterization as a carcinogen  in  the mid 20th century  (Nettleship, Henshaw and Meyer  1943), EC has been found ubiquitously  in almost all wines and spirits (Canas, et al. 1989; Coulon, et al.  2006). Although the actual health  implications of EC consumption by humans has yet to be definitively  established, the Canadian government has imposed a legal limit (30 µg/L) and the US FDA has imposed a  voluntary  limit  (15 µg/L) on  the EC  content of wines  (Butzke and Bisson 1998).  Furthermore,  several  agencies, including the National Institute of Health’s National Toxicology Programme and the University  of California Davis Department of Enology and Viticulture, have published preventative action manuals  for limiting the EC content of wines and spirits (Butzke and Bisson 1998).     Despite the mandatory and voluntary  limits imposed on the EC content in food and beverages,  its  presence  continues  to  be  a  pervasive  problem.  A  recent  survey  by  our  group  found  that  of  20  randomly chosen, commercially available wines from six wine producing countries, 14 and 17 exceeded  the  Canadian  and  US  limits,  respectively  (Coulon,  et  al.  2006).  It  is  therefore  obvious  that  current  methods  for  EC  reduction  in wine  are  largely  ineffective. Methods  currently  suggested  to  reduce  EC  include  agricultural  practices  to  control  grape must  arginine  content  (Butzke  and  Bisson  1998),  the  addition of  lyophilized acid urease preparations  to wine  (Kodama, et al. 1994; Ough and Trioli 1988),  and, to a lesser extent, the metabolic engineering of CAR1 deficient yeast strains was exploited to limit  EC in Sake (Kitamoto, et al. 1991; Park, Shin and Woo 2001; Yoshiuchi, Watanabe and Nishimura 2000).     Until  recently,  the  previously  mentioned  methods  were  the  only  methods  of  EC  reduction  available to winemakers. However,  in 2006 our group developed a functionally enhanced strain of the  popular industrial wine yeast 522 that is capable of reducing EC levels by 89% (Coulon, et al. 2006). This  result, which  is a consequence of the constitutive expression of an otherwise  inactive urea amidolyase  (DUR1,2) gene, far exceeds any other method of EC reduction and does not lengthen production time or      83    incur any additional costs to winemakers/consumers. The metabolically enhanced 522 strain contains a  constitutively expressed DUR1,2 gene, no antibiotic resistance marker genes or foreign DNA and is thus  non‐transgenic and  ‘self‐cloned’. This  yeast has  received approval  from  the  FDA, Health Canada, and  Environment Canada for commercial wine production.      4.2  Integration of the DUR1,2 cassette into the genomes of Sake yeast strains K7 and K9 yielded the  functional urea degrading Sake yeasts K7EC‐ and K9EC‐    S. cerevisiae produces wine with varying amounts of residual urea, mainly due to NCR of its urea  amidolyase  encoding  gene  (DUR1,2  ‐  YBR208C);  synthesis  of  the  urea  degrading  urea  amidolyase  enzyme is thus prevented and urea is excreted in to the wine (Kodama, et al. 1994; Monteiro and Bisson  1991; Monteiro, Trousdale and Bisson 1989; Ough, et al. 1990; Ough, et al. 1991; Stevens and Ough  1993). The successful integration of the DUR1,2 cassette (Figure 15) (Coulon, et al. 2006) into the URA3  locus of the popular Sake yeast strains K7 and K9 yielded the urea degrading strains K7EC‐ and K9EC‐.  In  addition to the URA3 flanking sequences required for homologous recombination, the DUR1,2 cassette  contains the S. cerevisiae DUR1,2 gene under the control of the constitutive S. cerevisiae PGK1 promoter  and terminator signals.    4.2.1  Genetic characterization of K7EC‐ and K9EC‐    The targeted integration of the DUR1,2 cassette into the URA3 locus (YEL021W) was confirmed  on Southern blots hybridized with both DUR1,2 and URA3 (Figures 16 and 17). These blots revealed that  K7EC‐ and K9EC‐ both contain a single copy of the ~9.1 kb  linear DUR1,2 cassette  integrated  into one of  their  URA3  loci.  Genomic  hybridization  also  confirmed  that  both  the  diploid  strains  K7EC‐  and  K9EC‐  retained a non‐disrupted URA3 locus, thus maintaining their uracil prototrophy. Nutritional prototrophy  is important for industrial yeasts as they are required to ferment substrates with highly variable nutrient  contents.    As the DUR1,2 cassette does not contain any antibiotic resistance markers, no positive selection  method was available  to  identify  transformants  that carried  the  integrated cassette; as a  result, urea  degrading yeasts were initially identified by colony PCR. Screening was completed with the assistance of      84    the  co‐transforming  plasmid  pUT332  which  contains  both  the  bla  and  Tn5ble  resistance  markers  (Ampicillin  and  phleomycin,  respectively);  pUT332  was  used  in  order  to  reduce  the  number  of  transformants  that  had  to  be  screened  by  colony  PCR.  By  successive  subculturing  on  non‐selective  media,  the  plasmid was  lost  and  this was  confirmed  on  Southern  blots  hybridized with  the  bla  and  Tn5ble genes  (Figure 18). Sake strains K7EC‐ and K9EC‐ are the  first metabolically engineered Sake yeast  strains  to  be  constructed  without  the  integration  of  antibiotic  resistance markers  or  E.  coli  vector  sequences,  thus making  them  suitable  for  commercialization;  these  strains will  be more  likely  to  be  accepted by consumers than strains containing antibiotic resistance marker genes or foreign DNA.    One  strand  of  the  integrated  DUR1,2  cassettes  in  K7EC‐  and  K9EC‐  was  sequenced.  Upon  comparison to previously published sequences (Coulon, et al. 2006), a single nucleotide difference was  observed  in  the  sequence of K9EC‐  (Table 8).  This  single nucleotide difference was  localized  to  the 5’  URA3  flanking  region of  the  cassette  and  is  likely due  to  a  genetic polymorphism between  the  Sake  strain K9 and  the wine  strain 522. No differences were observed  in  the DUR1,2  coding  region or  the  PGK1 promoter and terminator signals.     Analysis of DUR1,2 expression during wine fermentation revealed that integration of the DUR1,2  cassette indeed relieves repression of DUR1,2 by NCR, as the PGK1 promoter is much stronger than the  inducible/repressible, NCR sensitive DUR1,2 promoter. DUR1,2 mRNA was approximately 10‐fold more  abundant in K7EC‐ and K9EC‐ than in their respective parent strains (Figure 20).    The global gene expression pattern of the engineered strain K7EC‐ as compared to the pattern of  the parent strain K7 was studied at 24 hours into Chardonnay must fermentation. Besides DUR1,2 (6.35‐ fold overexpression), nine genes were affected ≥ 4‐fold; thus, it is evident that integration of the DUR1,2  cassette into the genome of S. cerevisiae K7 had a minimal effect (0.15% change) on the transcription of  the 5795 ORFs (4692 verified and 1103 uncharacterized, SGD, March, 2008) in the yeast cell. One gene  of interest that was upregulated ≥ 4‐fold in K7EC‐ (Table 10) was HAC1 (4.23‐fold), a transcription factor  involved  in the unfolded protein response (Cox and Walter 1996, Mori 1996, Nikawa, et al. 1996); this  response  is  likely  needed  to  support  the  increased  translation  and  folding  of DUR1,2p  constitutively  expressed  from  the  strong  PGK1  promoter;  indeed,  HAC1  was  also  upregulated  (5.67‐fold)  in  the  engineered  strain  K7D3 which  contains  the DUR3  gene  under  the  control  of  the  PGK1  promoter  and      85    terminator signals (Table 17). The data also suggests that no metabolic pathways were affected by the  presence of the DUR1,2 cassette integrated into K7EC‐; however, three (RME1, SSP1, SDS3) of the seven  genes  downregulated  in  K7EC‐  are  involved  in  different  aspects  of meiosis/sporulation  (Table  10).  As  sporulation can be triggered by nutrient deficiency (Malone 1990),  it  is reasonable that the  integration  of the DUR1,2 cassette, which results in constitutive utilization of urea as a nitrogen source, may cause  yeast cells to alter or delay their response to nutrient limitation.    4.2.2  The Sake yeasts K7EC‐ and K9EC‐ conduct efficient alcoholic fermentations     As measures  of  substantial  equivalence,  fermentation  rates,  glucose/fructose  utilization  and  ethanol  production  was  evaluated  in  the  metabolically  engineered  strains  K7EC‐  and  K9EC‐.  Prior  to  commercialization, engineered yeasts must obtain approval from regulatory agencies such as the FDA,  Health  Canada,  and  Environment  Canada.  In  North  America,  approval  is  granted  on  the  basis  of  substantial  equivalence, which means  that  should  a  foodstuff  from  a  genetically modified  organism  (GMO) be proven to be ‘as safe as’ the food produced by traditional means, then both should be treated  equally (Kessler 1992). Proof of substantial equivalence and safety for a genetically engineered organism  generally  requires  detailed  data  regarding  the  organism’s  genotype,  phenotype,  transcriptome,  proteome, and metabolome.     In  both  Chardonnay  must  and  Sake  mash,  K7EC‐  and  K9EC‐  conducted  efficient  alcoholic  fermentations (Figures 21 and 22), comparable to those of the parental strains K7 and K9; fermentations  were  completed  in  ~300  hours  in Chardonnay wine  and  ~500‐600 hours  in  Sake wine.  Furthermore,  residual glucose and  fructose  in Sake wine produced with K7EC‐ and K9EC‐  (0.05, 0.010 g/L glucose and  0.72, 0.30 g/L fructose, respectively) and the parental strains K7 and K9 were comparable (0.06, 0.08 g/L  glucose and 0.65, 0.31 g/L fructose, respectively). All four yeasts produced similar amounts of ethanol in  Sake wine (Table 11).      4.2.3 The Sake yeasts K7EC‐ and K9EC‐ reduce EC poorly in Chardonnay wine, yet efficiently in Sake wine     The  two  functionally  enhanced  Sake  yeasts K7EC‐  and K9EC‐ were  evaluated  for  their  ability  to  reduce EC content  in both Chardonnay wine and Sake wine.  In Chardonnay wine, K7EC‐ and K9EC‐ were      86    ineffective in reducing EC; K7EC‐ and K9EC‐ reduced EC by 30% and 0%, respectively (Table 12). In contrast,  K7EC‐ and K9EC‐ both effectively reduced EC by 68% in Sake wine (Table 13).     During the course of Sake fermentations, yeast cells are exposed to a substantially different set  of environmental conditions and stresses than those experienced during wine fermentation.  Specialized  Sake  yeast  strains have been  selected  for  the production of  Sake wine  (Shobayashi, et al. 2007) and  fermentation conditions undoubtedly play a role in Sake yeast metabolism. While several environmental  parameters may play  a  role  in  explaining  the  ineffective  EC  reduction of  Sake  strains  K7EC‐  and  K9EC‐  during Chardonnay wine making (Tables 12 and 13), the effect of yeast available nitrogen on native NCR  controlled genes is likely the predominant factor.      The type and quantity of yeast assimilable nitrogen (YAN) in Sake mash is substantially different  from grape must. Grape must of almost all varietals is usually high in free arginine and proline (Kliewer  1970). In contrast, Sake mash is rich in polypeptides that form structures known as protein bodies (PB)  (Kizaki, et  al. 1991).    These protein bodies, which  fall  into  two major  categories  (PB‐I  and PB‐II),  are  primarily composed of prolamins and glutelins, respectively (Hashizume, et al. 2006). During the course  of the fermentation yeast draw on the pool of free amino acids which is replenished by degradation of  PB by koji‐provided acid carboxypeptidases and acid proteases  (Hashizume, et al. 2006;  Iemura, et al.  1999a; Iemura, et al. 1999b). This controlled release process, which may  limit the  large scale  induction  of certain amino acid catabolic enzymes, functions to prevent nitrogen exhaustion (Wu, et al. 2006) and  subsequent TOR mediated transcriptional reprogramming leading to the de‐repression of NCR sensitive  genes (Cooper 2002; Hofman‐Bang 1999).    In contrast to the nitrogen available in Sake mash, yeasts draw on the large pool of free amino  acids in grape must to create biomass during fermentation. However, at approximately 2‐4 days into the  fermentation, cells  stop dividing and enter a  stationary growth phase  in which  they  ferment actively.  While there is new evidence to suggest that ethanol stress may play a role transitioning into stationary  phase (Marks, et al. 2003; Marks, et al. 2008), the general consensus in the literature, however, is that  nitrogen exhaustion  is  the primary  reason why  yeast  cells enter  into  stationary phase  (Hauser, et al.  2001; Rossignol, et al. 2003; Shobayashi, et al. 2007; Wu, et al. 2006). Nitrogen exhaustion, which  is  detected by the TOR pathway, is accompanied by large scale transcriptional changes leading to the de‐     87    repression of genes  involved  in the catabolism of poor nitrogen sources,  including urea (Cooper 2002;  Hofman‐Bang 1999). Interestingly, the second most abundant amino acid in grape must (Kliewer 1970),  proline,  is not metabolized during fermentation. Proline catabolism requires molecular oxygen and, as  fermentations  are  anaerobic, proline  cannot be utilized  (Ingledew, Magnus  and  Sosulski 1987). Thus,  after  nitrogen  exhaustion  shifts  cells  into  stationary  phase,  the  high  concentration  of  proline  in  the  ferment  likely  reinforces,  sustains and  strengthens  the  transcriptional  reprogramming associated with  NCR de‐repression.      Of particular interest to EC reduction is the de‐repression of the NCR sensitive urea amidolyase  encoding gene DUR1,2.  Induction of DUR1,2 allows yeast  to degrade urea  in an  increasingly nitrogen  scarce environment. Thus, during  the  later  stages of wine  fermentation, DUR1,2  can be  induced and  maintained throughout the fermentation as long as yeast are starved for nitrogen. As a result, parental  strains produce less EC overall since they tend to induce DUR1,2 at the middle/end of the fermentation,  thereby  reducing  the  effectiveness  of  functionally  enhanced  clones  (Table  12). During  Sake  brewing  however, no nitrogen exhaustion  is experienced by yeast cells and DUR1,2, along with  the other NCR  sensitive genes, remains largely repressed, resulting in higher absolute EC values and an increase in the  effectiveness of  the metabolically enhanced yeasts containing  the constitutive DUR1,2 cassette  (Table  13). Data obtained from global gene expression studies during grape wine and Sake wine fermentations  confirm that, during grape wine fermentation, the native DUR1,2 and DUR3 genes are both  induced at  the transition  into stationary phase (Rossignol, et al. 2003); this  is consistent with nitrogen exhaustion  and subsequent transcriptional reprogramming inducing stationary phase. In Sake wine making, DUR1,2  and DUR3 are transcriptionally inactive throughout the course of fermentation (Wu, et al. 2006), which  is consistent with an adequate nitrogen supply.    Another difference between grape must and rice mash concerns osmotic stress. As a result of  the  continuous  saccharification  process  during  Sake  brewing,  (Figure  4),  yeasts  are  subjected  to  substantially  less osmotic stress during Sake  fermentation than during wine  fermentation where all of  the sugars are present at  inoculation  (Takagi, et al. 2005; Wu, et al. 2006). This reduced stress during   Sake  fermentations  is  thought  to play a  role  in  the ability of Sake yeasts  to produce up  to 20%  (w/v)  ethanol  in  Sake wine  (Shobayashi,  et  al.  2007).  During  osmotic  stress,  yeast  cells  activate  the  ‘high  osmolarity growth’ (HOG) pathway which allows cells to survive osmotic stress via the biosynthesis and      88    intracellular retention of small molecules such as  trehalose and glycerol  (Westfall, Ballon and Thorner  2004).  These  molecules  help  to  offset  the  osmotic  pressure  created  by  the  high  concentration  of  extracellular  sugars  (often  20‐30% w/v,  equimolar  amounts  of  glucose  and  fructose  in  grape must),  which would  otherwise  crenate  the  cells.  Although  S.  cerevisiae  has  evolved  to  tolerate  substantial  osmotic  stress,  induction  and  maintenance  of  tolerance  mechanisms  requires  significant  energy  expenditure that could otherwise be devoted to biomass creation and other metabolic processes (Wu,  et  al.  2006).  For  Sake  yeast, which  has  evolved  in  a  niche  of markedly  reduced  osmotic  stress,  the  osmotic shock  in grape must may be a  factor  in explaining  the poor EC  reduction of engineered Sake  strains during wine making. Specifically,  the yeast  stress  response  tends  to down‐regulate  translation  globally,  which  in  turn  could  decrease  the  levels  of  DUR1,2p  (Cooper  2002;  Hauser,  et  al.  2001;  Rossignol, et al. 2003).     Finally, the addition of koji to Sake fermentations plays an  important role  in the availability of  some vital auxiliary factors. One such factor is ergosterol, a cholesterol derivative compound present in  yeast cell membranes that is vital for ethanol tolerance (Inoue 2000). Ethanol causes rigidity of the cell  membrane  and  induces  fatal  cracks;  ergosterol,  like  cholesterol,  functions  to  decrease  the  packing  density of membrane phospholipids and  increase membrane fluidity, thus counteracting the effects of  ethanol  (Inoue  2000).  Yeast  cells  require  oxygen  to  synthesize  ergosterol  (Jahnke  1983),  and  during  anaerobic  fermentations  this dependency on ergosterol may be  a  limiting  factor on  cell  viability  and  ethanol  production.  In  fact,  wine  fermentations  are  often  oxygenated  briefly  prior  to  or  during  fermentation  in  order  to  facilitate  the  production  of  ergosterol  (Rossignol,  et  al.  2003).  In  Sake  fermentations however, ergosterol is supplied primarily from koji which are cultured aerobically (Wu, et  al.  2006),  and  as  such  Sake  yeast  may  have  evolved  to  be  dependent  on  this  external  source  of  ergosterol.  In  fact,  Sake  yeast  strains were  shown  to  induce  all  of  the  ergosterol  biosynthetic  genes  during the course of Sake fermentation thus indicating an unexpected shortage of ergosterol despite the  presence of koji derived ergosterol (Wu, et al. 2006).  Thus, when functionally enhanced Sake yeasts are  used to ferment grape must which is not oxygenated, the lack of ergosterol contribution from koji may  play a role in a reduction in cell health, viability and EC reduction.     Given the observed differences in the EC reduction of the functionally enhanced yeasts K7EC‐ and  K9EC‐ in grape must and Sake mash, it is imperative that, in order to gather the most relevant data, the      89    functionality  of  engineered  yeasts  be  tested  in  their  niche  environments;  this  should  prevent  the  identification of false negatives and will expedite commercialization of new strains.    4.3  Constitutive expression of the urea permease, DUR3, in yeast cells is a viable alternative method  to reduce EC in fermented alcoholic beverages    The  most  efficient  metabolically  enhanced  wine  yeast  522EC‐,  which  contains  the  DUR1,2  cassette integrated into the URA3 locus, reduced EC by 89% in Chardonnay wine (Coulon, et al. 2006). In  order  to  further  improve  EC  reduction,  we  created  yeast  strains  that  acted  as  ‘urea  sponges’,  reabsorbing any urea secreted during fermentation or urea native to the must. This goal was completed  through the constitutive expression of the urea permease DUR3 (YHL016C), a NCR sensitive gene native  to S. cerevisiae (Cooper and Sumrada 1975; ElBerry, et al. 1993; Sumrada, Gorski and Cooper 1976). Due  to the abundance of other high quality nitrogen sources during primary fermentation, there is no need  for yeast cells to actively  import urea, thus the urea transporter  is transcriptionally silenced. Given the  success of the constitutively expressed DUR1,2 cassette previously characterized by our group (Coulon,  et  al.  2006),  we  chose  to  utilize  similar  methodology  in  order  to  transcriptionally  activate  and  constitutively express DUR3 in wine and Sake yeasts.     4.3.1  Construction of a linear PGK1p‐DUR3‐PGK1t‐kanMX cassette for integration into the TRP1 locus  of wine and Sake yeasts    Although  constitutive  expression  of DUR3  could  have  been more  easily  achieved  through  an  episomal plasmid expression system, creation of a linear cassette for integration into the yeast genome  was a more attractive option. The main problem with a plasmid borne system is that without a positive  selection pressure, such as the addition of antibiotics to the must, plasmid loss in S. cerevisiae generally  occurs within 3‐5 generations (Jones, Pringle and Broach 1992) thereby reverting engineered strains to  wild  type.  Coupled  with  the  nutrient  prototrophy  of  industrial  yeasts,  positive  selection  during  winemaking  is extremely  impractical as grape must/rice mash  is essentially a  ‘rich’ medium  for yeast  growth. Furthermore, addition of antibiotics  to  the  fermentation substrate  is undesirable and negates  the  goal  of  creating  ‘self‐cloned’  yeast  strains. Additionally,  S.  cerevisiae  is  highly  amenable  to  gene      90    uptake  and  integration  via  homologous  recombination  (Ausubel,  et  al.  1995;  Griffiths  et  al.,  2005),  making precise manipulation, replacement, and deletion of genes in the genome possible.    In order to create a linear cassette for the constitutive expression of DUR3 (Figure 25), the DUR3  ORF was placed under the control of the S. cerevisiae PGK1 promoter and terminator signals (Figure 10).  PGK1 (YCR012W) encodes 3‐phosphoglycerate kinase (EC 2.7.2.3) which is a key enzyme in the glycolytic  pathway responsible for transferring the acyl phosphate from 1,3‐bisphosphoglycerate to ADP thereby  creating one molecule of energy rich ATP (Blake 1981; Hitzeman 1980; Lam 1977). Being a key enzyme in  the ubiquitous process of glycolysis, PGK1, while technically inducible, is constitutively expressed so long  as  yeast are  grown  in  the presence of glucose  (Lam 1977). During both wine and  Sake  fermentation  glucose is abundant and cells never experience carbon exhaustion (Marks, et al. 2008; Wu, et al. 2006).  Additionally, as glycolysis is essential during fermentation (Lam 1977), placing DUR3 under the control of  PGK1 ensured constitutive expression throughout the course of fermentation.     As this section of the research was to be a proof of concept study only, the antibiotic resistance  marker  kanMX was  used  in  order  to  simplify  the  selection  of  positive  clones  (Figure  11). A  positive  selection marker allowed us to distinguish DUR3 clones from untransformed cells thus eliminating the  need  for  costly  and  time  consuming  colony  PCR  selection. While  this  approach was  suitable  for  this  study,  it will be necessary  to construct an antibiotic resistance  free cassette similar  to  that of DUR1,2  when DUR3 strains are developed for commercial purposes.     In order  to  target  the  linear DUR3  expression  cassette  to  a  specific  locus  in  the  genome  the  PGK1p‐DUR3‐PGK1t‐kanMX  cassette was  flanked  on  either  side with  300  bp  of  TRP1  homology  and  successfully integrated into the TRP1 locus (Figure 13). While homologous recombination in S. cerevisiae  is  possible  in  laboratory  strains with  as  few  as  40  nucleotides  of  flanking  homology  (Baudin  1993;  Manivasakam  1995),  we  used  300  homologous  nucleotides  on  either  end  of  the  DUR3  cassette  to  ensure efficient and specific integration into the TRP1 locus. Shorter regions of homology increases the  likelihood  that  strain  sequence polymorphisms will  reduce  integration efficiency. Efficiencies  increase  30‐ to 50‐fold when flanking sequences of several hundred base pairs in length are used (Wach 1996).         91    While the DUR3 cassette could have been integrated into any locus, the TRP1 locus was chosen  for  a  number  of  reasons.  Firstly,  TRP1  is  a  well  characterized  and  common  auxotrophic  marker  (Hampsey 1997; Mortimer 1966; Stolz 1998). Secondly, TRP1 is closely located (1 cM) to the centromere  of chromosome four (Mortimer 1966) and is thus highly stable, as genetic recombination during meiosis  is  proportional  to  a  locus’  distance  from  the  centromere  (Griffiths,  et  al.,  2005).   As  a  result  of  this  genetic stability, engineered strains would be suitable for active dry yeast production and industrial use.    Prior to  integration  into test strains, the DUR3 cassette  in the plasmid pUCMD was sequenced  (single strand) to confirm its structure; analysis revealed that the plasmid contained all the desired DNA  fragments in the correct order and orientation (Figure 14). Furthermore, sequencing data revealed nine  single  nucleotide  changes  along  the  length  of  cassette  (Table  14  and  Figure  23);  however,  all  four  mismatches in the DUR3 coding region were silent and the deduced DUR3p was identical in amino acid  sequence  and  length  to  that  published  for  DUR3p  on  SGD.  Five  of  the  nine  base  pair mismatches  localized  to  the PGK1 promoter;  four were C  to T  conversions. Given  that  the  in  silico  sequence was  assembled from single pass, single strand sequencing, the most logical explanation for the mismatches is  sequencing error. This conclusion is supported by the fact that none of the identified mismatches in the  PGK1  promoter  were  previously  reported  in  the  utilization  of  the  PGK1  promoter  for  constitutive  expression  (Coulon,  et  al.  2006;  Husnik,  et  al.  2006;  Volschenk,  et  al.  1997).  Furthermore,  during  characterization of the wine yeast ML01 (Husnik, et al. 2006), two mismatches that were found  in one  copy  of  the  PGK1  promoter were  absent  in  the  other  copy  (Husnik,  et  al.  2006),  indicating  that  the  sequence of  the PGK1 promoter may be prone  to  sequencing errors. Given  that  the  integrated DUR3  cassette was  fully functional, further sequencing of the cassette to  identify bona fide mismatches was  deemed non‐essential.     4.3.2    Integration of  the DUR3  cassette  into  the  genomes of K7, K7EC‐  , 522, and 522EC‐  yielded  the  functional urea transporting yeasts K7D3, K7EC‐D3, 522D3,  and 522EC‐D3     Like DUR1,2, DUR3 is subject to transcriptional repression by NCR during fermentation, resulting  in  the  inability  of  S.  cerevisiae  to  re‐absorb  excreted  urea  in  the  presence  of  good  nitrogen  sources  (ElBerry, et al. 1993; Hofman‐Bang 1999). This inability to absorb excreted urea is a contributing factor in  the  production  of wines with  high  residual  urea,  thus  leading  to  high  EC.  In  order  to  constitutively      92    express DUR3  throughout  fermentation,  the DUR3  cassette  (Figure 25) was  integrated  into  the  TRP1  locus of K7, K7EC‐  , 522, and 522EC‐, which yielded the functionally enhanced strains K7D3, K7EC‐D3, 522D3,  and  522EC‐D3  (Table  16).  The  enhanced  strains  were  genetically,  phenotypically,  and  functionally  characterized.    4.3.2.1  Integration of the DUR3 cassette into the genomes of K7D3, K7EC‐D3, 522D3, and 522EC‐D3 results  in constitutive expression of DUR3. Correct  integration of the DUR3 cassette  into the TRP1  locus was  confirmed by Southern blots probed with both DUR3 and TRP1 fragments. S. cerevisiae strains K7D3, K7EC‐ D3,  522D3,  and  522EC‐D3 were  all  shown  to  contain  a  single  copy  of  the  ~6.5  kb  linear DUR3  cassette  integrated  in  their TRP1  loci  (Figure 26). Blotting also  confirmed  that  the diploid  strains K7D3, K7EC‐D3,  522D3, and 522EC‐D3 retained a non‐disrupted TRP1 locus, thus maintaining their tryptophan prototrophy  and wild type phenotype.    In  the DUR3 cassette, expression of DUR3  is controlled by  the PGK1 promoter and  terminator  signals.  As  was  previously  observed  (Section  4.2.1)  during  the  characterization  of  DUR1,2  cassette  clones, PGK1  is a strong promoter capable of driving constitutive expression under NCR conditions. To  confirm  the  constitutive expression of DUR3  from  the  integrated DUR3 cassette, a northern blot was  hybridized with a DUR3 probe; blots of K7, K7EC‐, K7D3, and K7EC‐D3 total RNA revealed strong expression  of DUR3 in K7D3 and K7EC‐D3 (Figure 29). RNA for the northern blot was isolated from cells 24 hours into a  fermentation of Chardonnay must; at 24 hours, quality nitrogen sources are still abundant and thus NCR  is still strong. Thus, under the control of the PGK1 promoter and terminator signals, DUR3 expression is  high in cells containing the integrated DUR3 cassette despite non‐inducing conditions. Quantification of  constitutive expression by qRT‐PCR revealed a 14‐ and 7‐fold  increase  in DUR3 mRNA at 24 hours  into  Chardonnay must  fermentations  in K7D3  and K7EC‐D3,  respectively  (Figure 30).  Furthermore, high  level  expression of both DUR1,2 and DUR3 was maintained in K7EC‐D3 (12‐ and 7‐fold, respectively), which had  both  the  DUR1,2  and  DUR3  cassettes  integrated  into  its  genome.  These  data  suggest  that  K7,  and  presumably  other  strains  of  S.  cerevisiae,  is  tolerant  to  altered,  high  level  expression  of  multiple  proteins. Coupled with data confirming the functionality of both DUR1,2p (Table 21) and DUR3p (Figure  31), this result validates the metabolic engineering of S. cerevisiae for polygenic traits. In strains in which  only  one  cassette  (DUR1,2  or  DUR3)  was  integrated,  constitutive  expression  of  that  gene  induced  expression of the other gene (Figure 30). These results indicate a certain amount of cross talk between      93    the  regulatory mechanisms  for DUR1,2 and DUR3. Presumably when cells are actively degrading urea  instead  of  exporting  it  (constitutive  expression  of  DUR1,2),  the  urea  degradation  intermediate,  allophanate, causes induction of DUR3. Allophanate is a known inducer of all DUR genes (Cooper 1982;  Cooper 2002; Cooper and Sumrada 1975; ElBerry, et al. 1993; Hofman‐Bang 1999; Uemura, Kashiwagi  and  Igarashi  2007; Whitney  and  Cooper  1972; Whitney,  Cooper  and Magasanik  1973)  and  exerts  its  effect  through  an  upstream  induction  sequence  that  binds  the  transcriptional  activators  DAL81  and  DAL82, and through an upstream activation sequence that binds the NCR GATA factor GLN3 (Hofman‐ Bang  1999).  Similarly, when  cells  are  actively  importing  urea  (constitutive  expression  of  DUR3),  the  increased  intracellular  urea  concentration  induces  DUR1,2  expression  such  that  the  intracellular  concentration of urea can be lowered before it becomes toxic.     The global gene expression patterns of the metabolically engineered strain K7D3 and the parent  strain  K7  were  studied  at  24  hours  into  Chardonnay must  fermentation.  Besides  DUR3  (36.95‐fold  overexpression),  seven  genes were affected  ≥ 4‐fold;  thus,  it  is evident  that  integration of  the DUR3  cassette into the genome of S. cerevisiae K7 had a minimal effect (0.1% change) on the transcription of  the 5795 ORFs (4692 verified and 1103 uncharacterized, SGD, March, 2008) in the yeast cell. One gene  that was  upregulated  ≥  4‐fold  in  K7D3  (Table  17) was  common  to  those  upregulated  in  the  DUR1,2  cassette containing engineered  strain K7EC‐  (Table 10); HAC1  (5.67‐fold) encodes a  transcription  factor  involved  in the unfolded protein response and  is  likely needed for  increased translation and folding of  DUR3 constitutively expressed from the strong PGK1 promoter. The data also suggests that no metabolic  pathways  were  affected  by  the  presence  of  the  DUR3  cassette  integrated  into  K7EC‐.  However,  integration  of  the  DUR3  cassette  into  K7D3  appears  to  have  had  some  effect  on  the  regulation  of  meiosis/sporulation, processes  controlled by nutrient deficiency  (Malone 1990). One gene  (FIG1) was  upregulated  and  one  gene  (TID3)  was  downregulated  in  K7D3  (Table  17);  both  are  involved  in  meiosis/sporulation. As  constitutive  expression of DUR3  results  in  increased nitrogen  availability  it  is  reasonable that the DUR3 cassette would exert some effect on sporulation gene regulation.    4.3.2.2    The  integrated  DUR3  cassette  results  in  enhanced  urea  uptake.  In  order  to  confirm  the  production of a  functional urea permease encoded by  the  integrated DUR3 cassette, and  to correlate  DUR3 constitutive expression with increased urea uptake, the uptake of radiolabelled urea by K7, K7EC‐,  K7D3, and K7EC‐D3 was studied. Under the conditions tested, the strains K7D3 and K7EC‐D3 were both highly      94    capable of importing 14C‐urea while K7 and K7EC‐ were unable to incorporate any appreciable amounts of  14C‐urea  (Figure 31). These data  indicate  that  integration of  the DUR3  cassette  is  responsible  for  the  constitutive expression of an active urea permease (DUR3p) under NCR conditions.     In  the  14C‐urea  uptake  assay,  the  strains  K7,  K7EC‐,  K7D3,  and  K7EC‐D3  were  cultured  in  an  ammonium sulphate minimal medium that results in repression of all NCR sensitive genes (Cooper 1982;  Hofman‐Bang 1999). In all of the strains (K7, K7EC‐, K7D3 and K7EC‐D3) transcription of DUR3 from its native  promoter was repressed due the presence of ammonium sulphate; however in strains K7D3 and K7EC‐D3,  DUR3 was expressed despite the presence of ammonium sulphate in the medium due to the presence of  the recombinant DUR3 cassette. Given that the conditions of fermentation are largely repressive to NCR  sensitive  genes,  the  efficient  uptake  of  14C‐urea  by  K7D3  and  K7EC‐D3  in  conditions  of  strong  NCR  is  indicative of efficient EC reduction.    Despite integration of the same DUR3 cassette, K7D3 and K7EC‐D3 exhibited different rates of 14C‐ urea uptake; K7D3 was able to  incorporate approximately 3‐fold more 14C‐urea than K7EC‐D3 (Figure 31).  Since there is no data or rationale to explain why K7D3 would be more efficient at urea uptake than K7EC‐ D3,  the most  logical  explanation  is  that  the  constitutive  expression  of  DUR1,2  in  K7EC‐D3 masks  the  accumulation of 14C‐urea. In order to counteract the influx of toxic urea, K7D3 must induce transcription  of DUR1,2 and,  in turn, produce functional DUR1,2p. DUR1,2  is known to be controlled at the  level of  transcription (Genbauffe and Cooper 1986; Genbauffe and Cooper 1991); enzyme activity correlates well  with transcript abundance and the half‐life of DUR1,2 transcripts is the same in both the presence and  absence of an inducer (Jacobs, Dubois and Wiame 1985). In the case of K7EC‐D3, DUR1,2 is constitutively  expressed and  the  functional enzyme  is  capable of degrading urea as  it  is  incorporated  into  the  cell.  Thus, both K7D3 and K7EC‐D3 likely incorporate urea with equivalent efficiencies, however, this conclusion  was  not  confirmed  by  experimental  data.  The  same  masking  phenomenon  was  observed  when  comparing the uptake of 14C‐urea by K7 and K7EC‐ (Figure 32).     4.3.2.3   The metabolically engineered yeasts K7D3, K7EC‐D3, 522D3, and 522EC‐D3 ferment at similar rates  and produce  similar  amounts of  ethanol  in Chardonnay  and  Sake wine. As measures of  substantial  equivalence, fermentation rate and ethanol production was evaluated  in the metabolically engineered  strains K7D3, K7EC‐D3, 522D3, and 522EC‐D3.  In both Chardonnay must and Sake mash, K7D3, K7EC‐D3, 522D3,      95    and  522EC‐D3  conducted  substantially  equivalent  alcoholic  fermentations  (Figures  35  and  36).  Furthermore, the amount of ethanol produced in Chardonnay and Sake wine by K7D3, K7EC‐D3, 522D3, and  522EC‐D3 was shown to be similar (Tables 18 and 19). These results indicate that the K7D3, K7EC‐D3, 522D3,  and 522EC‐D3 strains are suitable for commercialization from a fermentation standpoint.    4.3.2.4  Variability of metabolically engineered yeasts to effectively reduce EC in Chardonnay wine and  in Sake wine. In order to assess the EC reduction potential of the engineered strains K7D3, K7EC‐D3, 522D3,  and 522EC‐D3, fermentations of both Chardonnay must and Sake rice mash were conducted after which  the EC content of the resultant wines was quantified. Consistent with prior observations of K7EC‐ (Section  4.2.2),  EC  reduction  by  K7EC‐  in  Chardonnay wine was  inefficient  (6.85%  ‐  Table  20)  due  to  low  EC  production by  the parental  strain K7. As expected, production of  Sake wine with K7EC‐  yielded highly  efficient  EC  reduction  (87%  ‐  Table  21).  Furthermore,  522EC‐  was  effective  at  reducing  EC  in  both  Chardonnay wine and Sake wine (81% and 84%, respectively ‐ Tables 20 and 21).     In Chardonnay wine, K7D3 and 522D3  reduced EC as efficiently as K7EC‐ and 522EC‐,  respectively  (Table 20); both K7D3 and K7EC‐ reduced EC by ~10% while 522D3 and 522EC‐ both reduced EC by ~81%. The  observed equivalency in EC reduction is likely a function of the need for cells to degrade urea once it is  internalized,  as  urea  is  toxic  to  cells  at  high  concentrations.  Presumably,  the  3.6‐fold  induction  of  DUR1,2  observed  in  K7D3  (Figure  30),  coupled  with  constitutive  urea  import,  is  responsible  for  the  strain’s ability  to  reduce EC as efficiently as K7EC‐. Despite  the  lack of gene expression data,  the same  rationale  is  likely responsible for the reduction of EC  in Chardonnay wine by 522D3. These data suggest  that K7 and 522 are highly sensitive to changes in the expression of genes involved in urea metabolism  and  that  engineering  yeasts  to  induce  large  scale  expression  changes  (greater  than  5‐fold) may  be  unnecessary.  Furthermore,  these  data  are  important  because  they  validate  the  application  of DUR3  constitutive expression  for  the  reduction of EC  in grape wine,  specifically  in wine derived  from musts  with high endogenous urea. In such cases, EC reduction would  likely be greatest  in wine fermented by  urea importing yeasts (DUR3 cassette) rather than urea degrading yeasts (DUR1,2 cassette).         In  Sake  wine,  K7D3  and  522D3  were  much  less  effective  in  reducing  EC  than  their  DUR1,2  counterparts K7EC‐ and 522EC‐ (Table 21). These results can be explained by the requirements for DUR1,2  expression  in S. cerevisiae (Figure 37). In a system analogous to the well characterized  lac operon  in E.      96    coli,  NCR  in  S.  cerevisiae  functions  to  force  cells  to  metabolize  the  most  energetically  favourable  nutrients. As such, expression of NCR regulated genes is highly dependent on two conditions being met:  the presence of an  inducer and the presence of an NCR associated transcriptional activator  like GLN3.  Indeed, expression of DUR1,2 was  shown  to be highly dependent on GLN3 under  induced  conditions  (Cooper, et al. 1990). Likewise, DUR1,2 expression on a proline medium (no NCR – GLN3 active) was very  low  without  an  inducer  present  (Cooper,  et  al.  1990).  As  discussed  previously  (Section  4.2.3),  fermentations  of  grape must  can  be  subject  to  nitrogen  exhaustion  and  thus  yeasts may  undergo  transcriptional de‐repression of NCR  sensitive genes  i.e. GLN3 becomes activated. During Chardonnay  must  fermentation K7D3 and 522D3 constitutively  imported urea  thus  fulfilling  the  inducer  requirement  for DUR1,2 expression. Upon nitrogen  limitation,  transcriptional  reprogramming  likely activated GLN3  and,  because  of  the  constitutive  presence  of  urea,  DUR1,2 was  induced  enabling  K7D3  and  522D3  to  degrade urea and reduce EC.  In contrast,  fermentations of Sake rice mash are not subject to nitrogen  exhaustion due  to  the presence of koji enzymes and  rice protein bodies. Although K7D3 and 522D3 still  constitutively  imported  urea  during  Sake  fermentations,  GLN3  remained  inactive  due  to  a  plentiful  supply  of  good  nitrogen.  Thus, without  satisfying  both  conditions  for  transcriptional  activation,  the  native  DUR1,2  genes  remained  repressed  in  K7D3  and  522D3  and  constitutively  absorbed  urea  likely  diffused back out of the cell, where  it could form EC  in the resultant Sake wine. Indeed, the facilitated  diffusion system for urea (DUR4), which allows leakage from the cell, is energy independent, insensitive  to NCR, and present in cells growing in the absence of any inducer (Cooper and Sumrada 1975).        97      Figure 37. Schematic representation of  inducible DUR1,2 expression during Chardonnay and Sake wine  fermentation by  a urea  importing  yeast  strain. During Chardonnay  fermentation partial  activation of  GLN3  coupled  with  the  constitutive  presence  of  the  inducer  allows  for  expression  of  DUR1,2  and  reduction of EC. Despite the constitutive presence of the inducer, constant NCR maintains repression of  DUR1,2 during Sake fermentations and constitutively imported urea is allowed to leak from the cell and  form EC.    While Figure 37 explains the most  likely reason for the  inability of K7D3 and 522D3 to reduce EC  efficiently  in  Sake  wine,  there may  be  other minor  issues.  Firstly,  yeast  growth  in  rice mash may  influence  the  degree  of  protein  mistargeting,  a  common  problem  for  membrane  protein  over  expression. Unlike  soluble proteins, which are  translated and kept  in  the cytosol, membrane proteins  must navigate the protein trafficking system which, consequently, is subject to a number of rate limiting  steps  (Reviewed  in Wagner, et al. 2006).  In eukaryotes, membrane protein production begins on  the  rough  endoplasmic  reticulum  (ER)  where  ribosomes  bound  to  ER  translate  proteins  which  are  immediately  translocated  into  the ER  lumen  (Wagner, et al. 2006). From  there, proteins can begin  to  undergo a variety of post‐translational modifications (most commonly glycosylation) and quality control  checks (Wagner, et al. 2006). Proteins  leave the ER by means of COPII coated vesicles that travel along  microtubules  through  the cis‐, mid‐ and  trans‐ Golgi apparatus before being excreted  into  the plasma  membrane via additional vesicles (Wagner, et al. 2006). Since protein trafficking machinery is limited in  its  speed  and  efficiency,  and  must  be  shared  with  other  essential  membrane  proteins,  excess  Minimal leakage of urea      98    overexpressed  protein  often  backs  up  (Wagner,  et  al.  2006).  Such  backups  are  often  dealt with  via  ubiquitin‐proteasome mediated  degradation  (Wagner,  et  al.  2006),  or,  as more  often  the  case  in  S.  cerevisiae, excess protein is diverted to the vacuole where it is either stored or degraded (Wagner, et al.  2006). Although expressed from a high copy plasmid rather than an integrated cassette, a recent study  of  the polyamine  transport  function of DUR3p reported significant amounts of over expressed DUR3p  localizing  to  the  vacuole  suggesting  a  high  degree  of  degradation  (Uemura,  Kashiwagi  and  Igarashi  2007).  This  result was  subsequently  confirmed  by  another  study  looking  at  optimizing  S.  cerevisiae  membrane protein over expression (Newstead, et al. 2007). Thus,  if cell growth  in rice mash  increases  DUR3p mistargeting  relative  to  growth  in  grape must,  or, more  likely,  if  cell  survival  in  Sake  wine  (perhaps due to high ethanol stress and low osmotic stress) requires the expression of a greater number  of membrane proteins,  the  amount of  functional DUR3p would be diminished,  thus  lowering  the  EC  reduction ability.     Finally, Sake brewing may change the primary physiological role of DUR3p. As noted previously  DUR3p  is primarily  an  active  transporter of urea, however  it has  also been  shown  to be  involved  in  boron  transport  (Nozawa, et al. 2006) and polyamine uptake  (Uemura, Kashiwagi and  Igarashi 2007).  While little is known about boron’s role in fermentation, polyamines have been shown to be crucial for  cell growth  (Tabor and Tabor 1999); polyamine  levels are elevated  in actively growing cells  (Monteiro  and  Bisson  1992).  Polyamines may  be  especially  important  for  growth  in  the  substantially  different  nutrient environment of rice mash; thus, during Sake brewing, an  increased need  for polyamines, and  perhaps boron, may decrease the amount of over expressed DUR3p available for urea uptake, thereby  lowering EC reduction ability.    4.3.2.5   The metabolically  engineered  yeasts K7EC‐D3  and 522EC‐D3 do not  reduce  EC more  effectively  than  K7EC‐  and  522EC‐  or  K7D3  and  522D3  in  either  Chardonnay  or  Sake wine.  The  effect  of  both  the  DUR1,2 and DUR3 cassettes on EC reduction was evaluated in Chardonnay and Sake wine. Strains K7EC‐D3  and 522EC‐D3 were not capable of producing Chardonnay or Sake wine with less EC than K7EC‐ and 522EC‐ or  K7D3 and 522D3 (Tables 20 and 21). Thus, it seems that constitutive expression of both DUR1,2 and DUR3  offers no  synergistic advantage over  the constitutive expression of either gene alone and  that,  in  the  case of metabolic engineering of S. cerevisiae to reduce EC in wine, it is prudent to limit engineering to      99    one  gene.  Yeast  strains  constitutively  expressing  only  a  single  gene  are  logistically  and  functionally  simpler, more cost effective, and can more expediently receive regulatory approval.    Data  from  this study suggested  that  in K7EC‐D3, which has both cassettes  integrated, high  level  expression  of  both DUR1,2  and DUR3 was maintained  (Figure  30),  and  that  both  cassettes  produce  functional enzymes (Figure 31). Moreover, urea degrading and urea importing yeast strains (integration  of  the DUR1,2  and DUR3  cassettes,  respectively)  each  reduce  EC  by  up  to  90%  (Tables  20  and  21);  however,  the  lack of synergistic EC  reduction  in strains constitutively co‐expressing DUR1,2 and DUR3  suggests the presence of confounding factors which must still be  investigated. One such factor may be  the induction of the native copies of DUR3 in the engineered strains at or near the end of fermentation.  If and when, at the end of fermentation, yeasts are subject to nitrogen exhaustion, the de‐repression of  NCR may result in the expression of DUR3; if the level of DUR3 expression from the native promoter is  equal or near equal to that of expression from the PGK1 promoter then it follows that integration of the  DUR3 cassette should not confer any advantage  in terms of urea uptake and EC reduction. To test this  hypothesis, the ability for yeasts to uptake urea under conditions of NCR de‐repression was examined  (Figures  33  and  34).  These  14C‐uptake  assays  were  completed  in  an  L‐proline  minimal  medium  to  simulate conditions of NCR de‐repression. Proline has  long been recognized as a poor quality nitrogen  source  for  yeast  (Hofman‐Bang  1999;  Ingledew,  Magnus  and  Sosulski  1987).  Under  de‐repressive  conditions, uptake of urea by the parental strain K7 and the engineered strains K7EC‐, K7D3, and K7EC‐D3  largely mimicked uptake under repressive conditions (Figures 31 and 33). Both K7D3 and K7EC‐D3, which  contained the integrated DUR3 cassette, were highly capable of incorporating 14C‐urea while the strains  that  did  not  contain  the  DUR3  cassette,  K7  and  K7EC‐, were  unable  to  incorporate  any  appreciable  amounts of urea  (Figure 33)  indicating  that constitutive expression of DUR3  from  the PGK1 promoter  and  terminator  signals  does  indeed  confer  an  advantage  to  engineered  yeast  cells  in  terms  of  urea  uptake. Furthermore, data obtained  from global gene expression  studies during grape wine and Sake  wine fermentations confirm that during grape wine fermentation DUR3 is only induced during nitrogen  exhaustion  (Rossignol,  et  al.  2003).  Moreover,  during  Sake  wine  making,  DUR3  is  transcriptionally  inactive throughout the course of  fermentation  (Wu, et al. 2006). Taken together, these data  indicate  that expression of DUR3 from  its native promoter  is not  likely to be a significant reason for the  lack of  synergistic EC reduction in engineered strains that constitutively import and degrade urea. Although the  de‐repressive conditions of the assay permitted full induction of native DUR3, the reason for poor urea      100    uptake in K7 and K7EC‐ may be the use of stationary phase cells for the assay. In order to express DUR3  from  its  native  promoter,  both  an  inducer must  be  present  (urea)  and  NCR must  be  de‐repressed  (Hofman‐Bang  1999). Despite  the  de‐repression  of NCR,  the  cells were  grown  in  the  absence  of  an  inducer  thus maintaining  low‐level DUR3 expression; only when  cells were exposed  to  14C‐urea were  both conditions for DUR3 expression met.     In  grape wine  and  Sake wine,  the major  contributor  to  EC  is  urea  along with  ethanol.  Post‐ fermentation  treatment of wine  and  Sake with  an  acid urease  resulted  in  significant  reduction of EC  (Ough and Trioli 1988); moreover, fermentation of Sake rice mash with CAR1 deficient yeast strains has  been  shown  to  produce  Sake wine with  no  detectable  urea  or  EC  (Kitamoto,  et  al.  1991;  Yoshiuchi,  Watanabe  and  Nishimura  2000).  These  studies  provide  strong  evidence  that  the  most  important  precursor  for  EC  is  urea  derived  from  CAR1  mediated  arginine  degradation;  however,  in  certain  circumstances,  such as wine  that has undergone a MLF, a  certain percentage of EC  could be derived  from alternate sources. While  this rationale would explain a  lack of synergistic EC reduction  in a wine  that had undergone MLF, it fails to explain the results of this study.    Besides  urea,  a  number  of  other  compounds  have  been  implicated  in  the  formation  of  EC.  Carbamyl phosphate and citrulline are known to react with ethanol and form EC (Matsudo 1993; Ough  1976b);  however, while  they  have  been  detected,  carbamyl  phosphate  and  citrulline  are  not  usually  found  in grape or Sake wine (Ough, Crowell and Gutlove 1988). Citrulline and carbamyl phosphate are  by‐products  of  arginine  metabolism  in  lactic  acid  bacteria  and  are  only  found  in  wines  that  have  undergone MLF;  in  lactic acid bacteria arginine  is metabolized via  the arginine deiminase pathway  in  which arginine  is degraded  to citrulline which  is  then degraded  to ornithine and carbamyl phosphate  (Liu, et al. 1994). Utilization of the metabolically engineered yeast strain ML01, which conducts parallel  alcoholic and malolactic fermentations without the addition of lactic acid bacteria (Husnik, et al. 2006),  could help reduce EC in certain applications. In stone fruit spirits such as fruit brandies, amygdalin from  the stone has been shown to degrade into cyanide during fermentation, the oxidized form (cyannate) of  which reacts with ethanol to form EC (Schehl 2007). Finally, a number of other compounds, such as N‐ carbamyl  α‐amino  acids, N‐carbamyl  β‐amino  acids,  and  allantoin  (by‐product of purine degradation)  have been shown to react with ethanol and  form EC  (Ough, Crowell and Gutlove 1988); however, the  contribution of these compounds to EC in alcoholic beverages has yet to be substantiated.       101    5  CONCLUSIONS    Amongst  other  reasons,  increased  awareness  surrounding  the  health  benefits  of  wine  consumption has caused  the popularity of wine  to grow worldwide. Wine, and  in particular  red wine,  has been shown to help prevent cardiovascular disease, dietary cancers, diabetes, hypertension, peptic  users, and macular degeneration, amongst others (Bisson 2002). Along with a growing appreciation for  the  benefits  of  alcoholic  beverage  consumption,  consumers  have  become  increasingly  conscious  of  consumption  risks  and  therefore  have  begun  to  demand  increasingly  safe  products.    This mindset  echoes beyond alcoholic beverages and  is highlighted by  the prevalence of  recent  food  safety  issues  including mad cow disease, pesticide usage, heavy metal contamination, E. coli O157:H7, GMO corn, and  other GMO  crops. Of particular  concern  to  this  study  is  the  increasing  recognition of  EC,  a naturally  occurring and well characterized carcinogen found in wine, Sake, and many other yeast fermented foods  and beverages. Indeed, winemakers and governments have become progressively more cognisant of the  widespread prevalence of the EC problem. Spearheaded by the FDA and the University of California at  Davis’ Department of Enology and Viticulture (Butzke and Bisson 1998), action manuals and data have  been  complied  concerning  a  wide  range  of  EC  reduction  techniques.  Until  recently,  none  of  the  techniques  employed  have  been  highly  effective  due  to  low  efficacy,  high  cost,  or  lengthened  production  time. However,  in  2006  our  group  developed  a  substantially  equivalent,  self‐cloned,  FDA  approved, industrial wine yeast (522EC‐) capable of reducing EC in Chardonnay wine by 89%. The amount  of EC in wines is proportional to residual urea, thus making EC reduction a matter of eliminating residual  urea at the end of alcoholic fermentation. S. cerevisiae strain 522EC‐ is capable of reducing EC by 89% due  to constitutive expression of urea amidolyase  (DUR1,2) which degrades urea  to ammonia and carbon  dioxide;  the  native  copy  of  this  gene  is  normally  transcriptionally  silent  in  fermentations  with  rich  nitrogen supplies thus resulting in wines with high residual urea and thus high EC .    Numerous industrial strains of S. cerevisiae exist around the world, each filling an environmental  niche. Arising from growth on different nutrient sources, strains have developed different fermentative  properties and, as such, are uniquely suited  to  fermentation of different substrates. However, EC  is a  pervasive problem in all types of alcoholic fermentation and, as a result, EC reduction is important in all  yeast strains. The self‐cloned Sake yeast strains K7EC‐ and K9EC‐ containing  the constitutively expressed  DUR1,2  cassette, were  ineffective when  tested  in wine making  trials;  however,  K7EC‐  and  K9EC‐ were      102    highly capable of EC reduction during Sake brewing trials. Thus,  it  is clear that certain yeast strains are  uniquely  adapted  to  specific  niches  and  that  they  are  not  optimally  functional  in  other  niches.  The  functionality of enhanced yeast strains must therefore be assessed in their native environment in order  to most accurately gauge performance and avoid the identification of false negatives.    In an effort to further reduce EC  levels beyond the capability of DUR1,2 technology, this study  was also focused on the creation of substantially equivalent, self‐cloned, industrial wine yeasts capable  of constitutive urea import. In order to create the industrial strains K7D3, K7EC‐D3, 522D3, and 522EC‐D3 the  S. cerevisiae urea permease (DUR3) was expressed under the control of the S. cerevisiae PGK1 promoter  and  terminator  signals  and  integrated  into  the  TRP1  locus  of  the  relevant  parental  strains.  DNA  sequencing confirmed that the recombinant cassette contained no deleterious mutations or amino acid  substitutions.  Furthermore,  Southern  blotting  confirmed  proper  cassette  integration  into  the  target  TRP1 locus. Constitutive expression of DUR3 was confirmed by northern blotting and quantified by qRT‐ PCR. Global gene expression analysis indicated that the integration of the DUR3 cassette had a minimal  impact on the transcriptome of S. cerevisiae and that no major metabolic pathways were affected by the  presence  of  the  DUR3  cassette.  Urea  import  studies  using  14C‐urea  indicated  that  DUR3  cassette  containing  strains  indeed  express  functional  DUR3p  and  that  this  functionality  is  preserved  when  DUR1,2  and DUR3  are  constitutively  co‐expressed.  Comparisons  of  small  scale  Chardonnay  and  Sake  wine making  demonstrated  that,  although  constitutive  co‐expression  of DUR1,2  and DUR3  does  not  yield  synergistic  EC  reduction  in  Chardonnay  or  Sake wine,  constitutive  expression  of DUR3  alone  is  capable of reducing EC as efficiently as constitutive expression of DUR1,2 in Chardonnay wine yet not in  Sake  wine.  Thus,  in  grape  wine  making  applications,  and  specifically  in  musts  which  contain  high  endogenous urea, DUR3 constitutive expression  is an  important, valuable, and alternative strategy  for  EC reduction. A provisional patent has been filed concerning the constitutive expression of DUR3  in S.  cerevisiae for the reduction of EC in grape wine. Furthermore, all of the DUR3 expressing strains can be  constructed by self‐cloned means, and thus would be more acceptable both worldwide due to their non‐ transgenic nature, and  in  countries  such as Germany and  Japan where  self‐cloned organisms are not  considered as GMO’s.            103    5.1  Future Directions    As a result of this study and previous work by our group (Coulon, et al. 2006), winemakers now  have highly effective methods  for EC reduction at their  fingertips; however, several topics still require  investigation.  Firstly,  in  order  to  facilitate  the  commercialization  of  yeast  strains  constitutively  expressing  DUR3,  a  cassette  must  be  developed  which  does  not  contain  any  antibiotic  resistance  markers.  Secondly,  experiments  should  be  done  in  order  to  confirm  which  of  the  potential  causes  discussed  in  Section 4.3.2.4  are  responsible  for  the  constitutive  expression of DUR3 not  reducing  EC  efficiently  in  Sake wine  as  it  does  in  Chardonnay wine;  answering  this  question may  allow  for  the  subsequent  adaptation  of DUR3  technology  to  Sake making.  Thirdly,  the  development  of  alternative  methods  of  EC  reduction would  prove  to  be  useful  in  further  reducing  EC  beyond  90%  thus making  wines and spirits safer for consumers. Moreover, as urea and EC production by different strains is highly  variable,  studies  into  the  molecular  mechanisms  behind  these  variations  will  aid  scientists  and  winemakers in engineering new strains or selecting natural strains which produce little or no EC. Finally,  it  is  imperative  that, before  the yeasts can be commercialized,  the proteome and metabolome of  the  metabolically  enhanced  yeast  strains  described  here  be  characterized  as  to  establish  substantial  equivalence.          104    REFERENCES  Ashby,  J.  (1991) Genotoxicity  data  supporting  the  proposed metabolic  activation  of  ethyl  carbamate  (urethane) to a carcinogen: the problem now posed by methyl carbamate. Mutat. Res. 260, 307‐ 308.  Ausubel,  F.M., Brent, R.,  Kingston, R.E., Moore, D.D.,  Seidman,  J.G.,  Smith,  J.A.  and  Struhl,  K.  (1995)  Short Protocols in Molecular Biology, 3rd ed., John Wiley & Sons, Inc., Boston, USA.  Baudin, A.  (1993) A simple and efficient method  for direct gene deletion  in Saccharomyces cerevisiae.  Nucl. Acids Res. 21, 3329‐3330.  Bisson, L.F. (2002) The present and future of the international wine industry. Nature 418, 696‐699.  Blake, C. C. (1981) Phosphoglycerate kinase. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 293, 93‐104.  Bond, U., Neal, C., Donnelly, D. and  James, T.  (2004) Aneuploidy and copy number breakpoints  in  the  genome of lager yeasts mapped by microarray hybridisation. Curr. Genet. 45, 360‐370.  Bruckner, R. and Titgemeyer, F.  (2002) Carbon catabolite  repression  in bacteria: choice of  the carbon  source and autoregulatory limitation of sugar utilization. FEMS Microbiol. Lett. 209, 141‐148.  Butzke,  C.E.  and  Bisson,  L.F.  (1998)  Ethyl  Carbamate  Preventative  Action  Manual.  2008.  (http://www.cfsan.fda.gov/~frf/ecaction.html).  Bysani, N., Daugherty,  J. R. and Cooper, T. G.  (1991) Saturation mutagenesis of  the UASNTR  (GATAA)  responsible for nitrogen catabolite repression‐sensitive transcriptional activation of the allantoin  pathway genes in Saccharomyces cerevisiae. J. Bacteriol. 173, 4977‐4982.  Canas, B. J., Havery, D. C., Robinson, L. R., Sullivan, M. P., Joe, F. L., Jr and Diachenko, G. W. (1989) Ethyl  carbamate levels in selected fermented foods and beverages. J. Assoc. Off. Anal. Chem. 72, 873‐ 876.      105    Cooper,  T.  G.  (1982)  Nitrogen  metabolism  in  S.  cerevisiae;  in  The  molecular  biology  of  the  yeast:  metabolism and gene expression. J.N. Strathern, E.W. Jones (ed.), pp. 39‐99, Cold Spring Harbor  Press, New York.  Cooper, T. G. (2002) Transmitting the signal of excess nitrogen in Saccharomyces cerevisiae from the Tor  proteins to the GATA factors: connecting the dots. FEMS Microbiol. Rev. 26, 223‐238.  Cooper,  T.  G.,  Ferguson,  D.,  Rai,  R.  and  Bysani,  N.  (1990)  The  GLN3  gene  product  is  required  for  transcriptional  activation  of  allantoin  system  gene  expression  in  Saccharomyces  cerevisiae.  J.  Bacteriol. 172, 1014‐1018.  Cooper, T. G. and Sumrada, R. (1975) Urea transport in Saccharomyces cerevisiae. J. Bacteriol. 121, 571‐ 576.  Coulon, J., Husnik, J. I., Inglis, D. L., van der Merwe, G. K., Lonvaud, A., Erasmus, D. J. and van Vuuren, H.  J.  J.  (2006) Metabolic  Engineering of  Saccharomyces  cerevisiae  to Minimize  the Production of  Ethyl Carbamate in Wine. Am. J. Enol. Vitic. 57, 113‐124.  Cox, J. S. and Walter, P. (1996) A Novel Mechanism for Regulating Activity of a Transcription Factor That  Controls the Unfolded Protein Response. Cell, 87, 391‐404.  Cox, K. H., Rai, R., Distler, M., Daugherty, J. R., Coffman, J. A. and Cooper, T. G. (2000) Saccharomyces  cerevisiae GATA sequences function as TATA elements during nitrogen catabolite repression and  when Gln3p is excluded from the nucleus by overproduction of Ure2p. J. Biol. Chem. 275, 17611‐ 17618.  Cox,  K.  H.,  Kulkarni,  A.,  Tate,  J.  J.  and  Cooper,  T.  G.  (2004)  Gln3  Phosphorylation  and  Intracellular  Localization  in Nutrient  Limitation  and  Starvation Differ  from  Those Generated  by Rapamycin  Inhibition of Tor1/2 in Saccharomyces cerevisiae. J. Biol. Chem. 279, 10270‐10278.  Dahl, G., Miller,  J.  and Miller,  E.  (1978)  Vinyl  carbamate  as  a  promutagen  and  a more  carcinogenic  analog of ethyl carbamate. Cancer Res. 38, 3793‐3804.      106    Dann, S. G. and Thomas, G. (2006) The amino acid sensitive TOR pathway from yeast to mammals. FEBS  Letters 580, 2821‐2829.  De Virgilio, C. and Loewith, R. (2006) The TOR signalling network from yeast to man. Int. J. Biochem. Cell  Bio. 38, 1476‐1481.  Dohmen, R. J. and Varshavsky, A. (2005) Heat‐inducible degron and the making of conditional mutants.  Methods Enzymol. 399, 799‐822.  Dunn, B., Levine, R. P. and Sherlock, G. (2005) Microarray karyotyping of commercial wine yeast strains  reveals shared, as well as unique, genomic signatures. BMC Genomics 6, 53‐74.  Eckhardt, T. (1978) A rapid method for the  identification of plasmid desoxyribonucleic acid  in bacteria.  Plasmid 1, 584‐588.  ElBerry, H. M., Majumdar, M. L., Cunningham, T. S., Sumrada, R. A. and Cooper, T. G. (1993) Regulation  of the urea active transporter gene (DUR3) in Saccharomyces cerevisiae. J. Bacteriol. 175, 4688‐ 4698.  Farrell A., E., Plevin R., J., Turner B., T., Jones A., D., O'Hare, M. and Kammen D., M. (2006) Ethanol Can  Contribute to Energy and Environmental Goals. Science 311, 506‐508.  Gancedo , J. M. (1992) Carbon catabolite repression in yeast. Eur. J. Biochem.  206, 297‐313.  Gatignol, A. (1987) Phleomycin resistance encoded by the ble gene from transposon Tn 5 as a dominant  selectable marker in Saccharomyces cerevisiae. Mol. Gen. Genetics 207, 342‐348.  Genbauffe,  F.  S.  and  Cooper,  T.  G.  (1991)  The  urea  amidolyase  (DUR1,2)  gene  of  Saccharomyces  cerevisiae. DNA Seq. 2, 19‐32.  Genbauffe,  F.  S.  and  Cooper,  T. G.  (1986)  Induction  and  repression  of  the  urea  amidolyase  gene  in  Saccharomyces cerevisiae. Mol. Cell. Biol. 6, 3954‐3964.  Gietz, R. D. and Woods, R. A. (2002) Transformation of yeast by  lithium acetate/single‐stranded carrier  DNA/polyethylene glycol method. Methods Enzymol. 350, 87‐96.      107    Goffeau, A., Barrell, B. G., Bussey, H., Davis, R. W., Dujon, B., Feldmann, H., Galibert, F., Hoheisel, J. D.,  Jacq, C.,  Johnston, M., Louis, E.  J., Mewes, H. W., Murakami, Y., Philippsen, P., Tettelin, H. and  Oliver, S. G. (1996) Life with 6000 Genes. Science 274, 546‐567.  Goode,  J.  (2005)  The  Science  of Wine:  From  Vine  to  Glass,  University  of  California  Press,  Berkley,  California.  Greig, D. and Travisano, M. (2003) Evolution: Haploid Superiority. Science 299, 524‐525.  Griffiths,  A.  J.  F.,  Miller,  J.  H.,  Suzuki,  D.  T.,  Lewontin,  R.  C.,  and  Gelbart,  W.  M.  (2005)   An Introduction to Genetic Analysis. Eight ed., W. H. Freeman, New York, USA.  Guengerich, F. P. and Kim, D. H. (1991) Enzymatic oxidation of ethyl carbamate to vinyl carbamate and  its role as an intermediate in the formation of 1,N6‐ethenoadenosine. Chem. Res. Toxicol. 4, 413‐ 421.  Guldener, U., Heck, S., Fielder, T., Beinhauer, J. and Hegemann, J. (1996) A new efficient gene disruption  cassette for repeated use in budding yeast. Nucl. Acids Res. 24, 2519‐2524.  Hampsey, M. (1997) A review of phenotypes in Saccharomyces cerevisiae. Yeast 13, 1099‐1133.  Han,  J.,  Lee,  J.,  Bibbs,  L.  and  Ulevitch,  R.  (1994)  A  MAP  kinase  targeted  by  endotoxin  and  hyperosmolarity in mammalian cells. Science 265, 808‐811.  Hashizume, K., Okuda, M., Sakurao, S., Numata, M., Koseki, T., Aramaki,  I., Kumamaru, T. and Sato, H.  (2006) Rice protein digestion by sake koji enzymes: comparison between steamed rice grains and  isolated protein bodies from rice endosperm. J. Biosci. Bioeng. 102, 340‐345.  Hauser,  N.,  Fellenberg,  R.,  Gil,  R.,  Bastuk,  S.,  Hoheisel,  J.  and  Perez‐Ortin,  J.  (2001) Whole  genome  analysis of a wine yeast strain. Comp. Funct. Genomics 2, 69‐79.  Herskowitz, I. (1988) Life cycle of the budding yeast Saccharomyces cerevisiae. Microbiol. Rev. 52, 536‐ 553.      108    Hitzeman, R. A. (1980) Isolation and characterization of the yeast 3‐phosphoglycerokinase gene (PGK) by  an immunological screening technique. J. Biol. Chem. 255, 12073‐12080.  Hofman‐Bang, J. (1999) Nitrogen catabolite repression in Saccharomyces cerevisiae. Mol. Biotechnol. 12,  35‐73.  Hughes, T., Roberts, C., Dai, H., Jones, A., Meyer, M., Slade, D., Burchard, J., Dow, S., Ward, T., Kidd, M.,  Friend, S. and Marton, M. (2000) Widespread aneuploidy revealed by DNA microarray expression  profiling. Nat. Genet. 25, 333‐337.  Husnik,  J.  I., Volschenk, H., Bauer,  J., Colavizza, D.,  Luo,  Z.  and  van Vuuren, H.  J.  J.  (2006) Metabolic  engineering of malolactic wine yeast. Metabol. Eng. 8, 315‐323.  Iemura,  Y.,  Takahashi,  T.,  Yamada,  T.,  Furukawa,  K.  and Hara,  S.  (1999a)  Properties  of  TCA‐Insoluble  peptides in Kimoto (traditional seed mash for sake brewing) and conditions for liberation of the  peptides from rice protein. J. Biosci. Bioeng. 88, 531‐535.  Iemura,  Y.,  Yamada,  T.,  Takahashi,  T.,  Furukawa,  K.  and Hara,  S.  (1999b)  Properties  of  the  peptides  liberated from rice protein in Sokujo‐moto. J. Biosci. Bioeng. 88, 276‐280.  Ingledew, W. M., Magnus, C. A. and Patterson, J. R. (1987) Yeast Foods and Ethyl Carbamate Formation  in Wine. Am. J. Enol. Vitic. 38, 332‐335.  Ingledew, W. M., Magnus, C. A. and Sosulski, F. W.  (1987)  Influence of Oxygen on Proline Utilization  During the Wine Fermentation. Am. J. Enol. Vitic. 38, 246‐248.  Inoue,T  .  (2000) Cloning and Characterization of a Gene Complementing  the Mutation of an Ethanol‐ sensitive Mutant of Sake Yeast. Biosci. Biotech. Biochem. 64, 229‐236.  Jacobs, E., Dubois, E. and Wiame, J. M. (1985) Regulation of urea amidolyase synthesis in Saccharomyces  cerevisiae, RNA analysis, and cloning of the positive regulatory gene DURM. Curr. Genet. 9, 333‐ 339.      109    Jahnke,  L.  (1983)  Oxygen  requirements  for  formation  and  activity  of  the  squalene  epoxidase  in  Saccharomyces cerevisiae. J. Bacteriol  155, 488‐492.  Jones  E.  W.,  Pringle  J.  R.,  Broach  J.  R.  (1992)  The  Molecular  and  Cellular  Biology  of  the  Yeast  Saccharomyces, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, USA.   Kessler, D. A. (1992) The safety of foods developed by biotechnology. Science 256, 1747‐1749.  Kinzy,  T.  G.  (1994)  Multiple  genes  encode  the  translation  elongation  factor  EF‐1  gamma  in  Saccharomyces cerevisiae. Nucl. Acids Res. 22, 2703‐2707.  Kitamoto, K., Oda, K., Gomi, K. and Takahashi, K. (1991) Genetic engineering of a sake yeast producing  no urea by successive disruption of arginase gene. Appl. Environ. Microbiol. 57, 301‐306.  Kizaki, Y., Inoue, Y., Okazaki, N. and Kobayashi, S. (1991) Isolation and determination of protein bodies  (PB‐I, PB‐II) in polished rice endosperm. J. Brew. Soc. Jpn. 86, 293‐298.  Kliewer, W.M. (1970) Free amino acids and other nitrogenous fractions  in wine grapes. J. Food Sci. 35,  17‐21.  Kodama, K. (1993) Sake‐brewing yeast; in The Yeasts, Rose, A. H. and Harrison, J. S. (eds.), pp. 129‐168,  Academic Press, London, United Kingdom.  Kodama, S., Suzuki, T., Fujinawa, S., de  la Teja, P. and Yotsuzuka, F.  (1994) Urea Contribution  to Ethyl  Carbamate Formation in Commercial Wines During Storage. Am. J. Enol. Vitic. 45, 17‐24.  Lam, K. B. (1977) Isolation and characterization of Saccharomyces cerevisiae glycolytic pathway mutants.  J. Bacteriol 130, 746‐749.  Leithauser, M. T.,  Liem, A.,  Stewart, B. C., Miller, E. C.  and Miller,  J. A.  (1990) 1,N6‐ethenoadenosine  formation,  mutagenicity  and  murine  tumor  induction  as  indicators  of  the  generation  of  an  electrophilic epoxide metabolite of  the closely  related carcinogens ethyl carbamate  (urethane)  and vinyl carbamate. Carcinogenesis 11, 463‐473.      110    Liu, S., Pritchard, G. G., Hardman, M. J. and Pilone, G. J. (1994) Citrulline Production and Ethyl Carbamate  (Urethane)  Precursor  Formation  From  Arginine  Degradation  by  Wine  Lactic  Acid  Bacteria  Leuconostoc oenos and Lactobacillus buchneri. Am. J. Enol. Vitic. 45, 235‐242.  Lohr, D., Venkov, P. and Zlatanova, J. (1995) Transcriptional regulation  in the yeast GAL gene family: a  complex genetic network. FASEB J. 9, 777‐787.  Maftahi, M., Gaillardin, C. and Nicaud,  J. M.  (1998) Generation of Saccharomyces cerevisiae deletants  and basic phenotypic analysis of eight novel genes from the  left arm of chromosome XIV. Yeast  14, 271‐280.  Malone, R. E. (1990) Dual regulation of meiosis in yeast. Cell, 61, 375‐378.  Manivasakam,  P.  (1995) Micro‐homology mediated  PCR  targeting  in  Saccharomyces  cerevisiae.  Nuc.  Acids Res. 23, 2799‐2800.  Marks, V. D., Sui, S.  J., Erasmus, D., van der Merwe, G., Brumm,  J., Wasserman, W. W., Bryan,  J., van  Vuuren, H. J. J. (2008) Dynamics of the yeast transcriptome during wine fermentation reveals a  novel fermentation stress response. FEMS Yeast Research 8, 35‐52.  Marks, V. D.,   van der Merwe, G., van Vuuren, H.  J.  J.  (2003) Transcriptional profiling of wine yeast  in  fermenting  grape  juice:  regulatory  effect of diammonium phosphate.  FEMS  Yeast Research 3,  269‐287.  Matsudo, T. (1993) Determination of ethyl carbamate in soy sauce and its possible precursor. J. Agricult.  Food Chem. 41, 352‐356.  Miklos, G.  G.  and  Rubin, G.  (1996)  The  Role  of  the Genome  Project  in Determining Gene  Function:  Insights from Model Organisms. Cell, 86, 521‐529.  Monteiro, F. F. and Bisson, L. F.  (1991) Amino Acid Utilization and Urea Formation During Vinification  Fermentations. Am. J. Enol. Vitic. 42, 199‐208.  Monteiro, F. F. and Bisson, L. F. (1992) Nitrogen Supplementation of Grape Juice. I. Effect on Amino Acid  Utilization During Fermentation. Am. J. Enol. Vitic. 43, 1‐10.      111    Monteiro,  F.  F., Trousdale, E. K.  and Bisson,  L.  F.  (1989) Ethyl Carbamate  Formation  in Wine: Use of  Radioactively Labeled Precursors to Demonstrate the Involvement of Urea. Am. J. Enol. Vitic. 40,  1‐8.  Mori,  K.  (1996)  Signalling  from  endoplasmic  reticulum  to  nucleus:  transcription  factor with  a  basic‐ leucine  zipper motif  is  required  for  the unfolded protein‐response pathway. Genes  to Cells 1,  803‐817.  Mortimer, R. K. (1966) Genetic mapping in Saccharomyces. Genetics 53, 165‐173.  Nasmyth, K. (1993) Regulating the HO endonuclease in yeast. Curr. Opin. Genet. Dev. 3, 286‐294.  National  Institutes  of  Health  National  Toxicology  Program.  (2004)  NTP  Technical  Report  on  the  Toxicology  and  Carcinogenesis  Studies  of Urethane,  Ethanol,  and Urethane/Ethanol  in B6C3F1  Mice (Drinking Water Studies). TR 510, 1‐351.  Newstead, S., Kim, H., von Heijne, G., Iwata, S. and Drew, D. (2007) High‐throughput fluorescent‐based  optimization of eukaryotic membrane protein overexpression and purification in Saccharomyces  cerevisiae. Proc. Natl. Acad. Sci. USA 104, 13936‐13941.  Nettleship, A., Henshaw, PS. and Meyer, HL.  (1943)  Induction of pulmonary  tumors  in mice with ethyl  carbamate (urethane). J. Natl. Cancer Inst. 4, 309‐319.  Nikawa, J., Akiyoshi, M., Hirata, S. and Fukuda, T. (1996) Saccharomyces cerevisiae IRE2/HAC1 is involved  in IRE1‐mediated KAR2 expression. Nucl. Acids Res. 24, 4222‐4226.  Nozawa, A., Takano, J., Kobayashi, M., von Wiren, N. and Fujiwara, T. (2006) Roles of BOR1, DUR3, and  FPS1  in boron  transport and  tolerance  in Saccharomyces cerevisiae. FEMS Microbiol. Lett. 262,  216‐222.  Ough,  C.  S.  (1976a)  Ethyl  carbamate  in  fermented  beverages  and  foods.  I. Naturally  occurring  ethyl  carbamate. J. Agric. Food Chem. 24, 323‐328.      112    Ough, C. S.  (1976b) Ethyl carbamate  in fermented beverages and foods.  II. Possible formation of ethyl  carbamate from diethyl dicarbonate addition to wine. J. Agric. Food Chem. 24, 328‐331.  Ough, C. S., Crowell, E. A. and Gutlove, B. R. (1988) Carbamyl Compound Reactions with Ethanol. Am. J.  Enol. Vitic. 39, 239‐242.  Ough, C. S., Crowell, E. A. and Mooney, L. A.  (1988) Formation of Ethyl Carbamate Precursors During  Grape Juice (Chardonnay) Fermentation. I. Addition of Amino Acids, Urea, and Ammonia: Effects  of Fortification on Intracellular and Extracellular Precursors. Am. J. Enol. Vitic. 39, 243‐249.  Ough, C. S., Huang, Z., An, D. and Stevens, D. (1991) Amino Acid Uptake by Four Commercial Yeasts at  Two  Different  Temperatures  of  Growth  and  Fermentation:  Effects  on  Urea  Excretion  and  Reabsorption. Am. J. Enol. Vitic. 42, 26‐40.  Ough,  C.  S.,  Stevens,  D.,  Sendovski,  T.,  Huang,  Z.  and  An,  D.  (1990)  Factors  Contributing  to  Urea  Formation in Commercially Fermented Wines. Am. J. Enol. Vitic. 41, 68‐73.  Ough, C. S. and Trioli, G. (1988) Urea Removal from Wine by an Acid Urease. Am. J. Enol. Vitic. 39, 303‐ 307.  Park, H., Shin, M. and Woo, I. (2001) Antisense‐mediated inhibition of arginase (CAR1) gene expression  in Saccharomyces cerevisiae. J. Biosci. Bioeng. 92, 481‐484.  Park,  K.,  Liem,  A.,  Stewart,  B.  C.  and Miller,  J.  A.  (1993)  Vinyl  carbamate  epoxide,  a major  strong  electrophilic, mutagenic  and  carcinogenic metabolite  of  vinyl  carbamate  and  ethyl  carbamate  (urethane). Carcinogenesis 14, 441‐450.  Pena‐Castillo,  L.  and Hughes, T.  (2007) Why Are There  Still Over 1000 Uncharacterized Yeast Genes?  Genetics 176, 7‐14.  Perez‐Ortin, J., Garcia‐Martinez, J. and Alberola, T. (2002) DNA chips for yeast biotechnology. The case of  wine yeasts. J. Biotechnol. 98, 227‐241.      113    Rossignol, T., Dulau, L.,  Julien, A. and Blondin, B.  (2003) Genome‐wide monitoring of wine yeast gene  expression during alcoholic fermentation. Yeast 20, 1369‐1385.  Salmon, J. and Barre, P. (1998) Improvement of Nitrogen Assimilation and Fermentation Kinetics under  Enological  Conditions  by  Derepression  of  Alternative  Nitrogen‐Assimilatory  Pathways  in  an  Industrial Saccharomyces cerevisiae Strain. Appl. Environ. Microbiol. 64, 3831‐3837.  Schehl, B.  (2007) Contribution of  the  fermenting  yeast  strain  to ethyl  carbamate generation  in  stone  fruit spirits. Appl. Microbiol. Biotech. 74, 843‐850.  Schlatter,  J.  and  Lutz, W.  K.  (1990)  The  carcinogenic  potential  of  ethyl  carbamate  (urethane):  risk  assessment at human dietary exposure levels. Food Chem. Toxicol. 28, 205‐211.  Shobayashi, M., Ukena, E., Fujii, T. and Iefuji, H. (2007) Genome‐wide expression profile of sake brewing  yeast under shaking and static conditions. Biosci. Biotechnol. Biochem. 71, 323‐335.  Smart, W. C., Coffman,  J. A. and Cooper, T. G.  (1996) Combinatorial  regulation of  the Saccharomyces  cerevisiae CAR1  (arginase) promoter  in  response  to multiple  environmental  signals. Mol. Cell.  Biol. 16, 5876‐5887.  Stevens, D. F. and Ough, C. S.  (1993) Ethyl Carbamate Formation: Reaction of Urea and Citrulline with  Ethanol in Wine Under Low to Normal Temperature Conditions. Am. J. Enol. Vitic. 44, 309‐312.  Stolz, L. E. (1998) INP51, a yeast inositol polyphosphate 5‐phosphatase required for phosphatidylinositol  4,5‐bisphosphate homeostasis  and whose  absence  confers  a  cold‐resistant phenotype.  J. Biol.  Chem. 273, 11852‐11861.  Sumrada,  R.,  Gorski,  M.  and  Cooper,  T.  (1976)  Urea  transport‐defective  strains  of  Saccharomyces  cerevisiae. J. Bacteriol. 125, 1048‐1056.  Tabor, H. and Tabor, C. W.  (1999) A Guide  to  the Polyamines. Seymour S. Cohen. Analytical Biochem.  274, 150. Oxford University Press, New York, USA.      114    Takagi, H.,  Takaoka, M.,  Kawaguchi, A.  and  Kubo,  Y.  (2005)  Effect  of  L‐Proline  on  Sake  Brewing  and  Ethanol Stress in Saccharomyces cerevisiae. Appl. Environ. Microbiol. 71, 8656‐8662.  Uemura,  T.,  Kashiwagi,  K.  and  Igarashi,  K.  (2007)  Polyamine  Uptake  by  DUR3  and  SAM3  in  Saccharomyces cerevisiae. J. Biol. Chem. 282, 7733‐7741.  van Vuuren, H. J. J., Daugherty, J. R., Rai, R. and Cooper, T. G. (1991) Upstream induction sequence, the  cis‐acting element required for response to the allantoin pathway  inducer and enhancement of  operation of the nitrogen‐regulated upstream activation sequence  in Saccharomyces cerevisiae.  J. Bacteriol. 173, 7186‐7195.  Vine, R. P., Harkness, E. M. and Linton, S. J. (2002) Winemaking: From Grape Growing to Marketplace,  Second ed., Kluwer Academic/Plenum, New York.  Volschenk, H., Viljoen, M., Grobler, J., Bauer, F., Lonvaud‐Funel, A., Denayrolles, M., Subden, R. E. and  Van Vuuren, H. J. J. (1997) Malolactic Fermentation in Grape Musts by a Genetically Engineered  Strain of Saccharomyces cerevisiae. Am. J. Enol. Vitic. 48, 193‐197.  Wach,  A.  (1996)  PCR‐synthesis  of  marker  cassettes  with  long  flanking  homology  regions  for  gene  disruptions in S. cerevisiae. Yeast 12, 259‐265.  Ward, A. (1992) Rapid analysis of yeast transformants using colony‐PCR. Biotechniques, 13, 350.   Wagner,  S.,  Bader,  M.  L.,  Drew,  D.  and  de  Gier,  J.  (2006)  Rationalizing  membrane  protein  overexpression. Trends in Biotech. 24, 364‐371.  Westfall, P. J., Ballon, D. R. and Thorner, J. (2004) When the stress of your environment makes you go  HOG wild. Science 306, 1511‐1512.  Whitney, P. A. and Cooper, T. G. (1972) Urea Carboxylase and Allophanate Hydrolase. Two components  of  adenosine  triphosphate:  urea  amidolyase  in  Saccharomyces  cerevisiae.  J.  Biol.  Chem.  247,  1349‐1353.      115    Whitney, P. A. and Cooper, T.  (1973) Urea carboxylase  from Saccharomyces cerevisiae. Evidence  for a  minimal two‐step reaction sequence. J. Biol. Chem. 248, 325‐330.  Whitney,  P.  A.,  Cooper,  T.  G.  and  Magasanik,  B.  (1973)  The  induction  of  urea  carboxylase  and  allophanate hydrolase in Saccharomyces cerevisiae. J. Biol. Chem. 248, 6203‐6209.  Wu, H., Zheng, X., Araki, Y., Sahara, H., Takagi, H. and Shimoi, H. (2006) Global gene expression analysis  of yeast cells during sake brewing. Appl. Environ. Microbiol. 72, 7353‐7358.  Yanisch‐Perron,  C.,  Vieira,  J.  and Messing,  J.  (1985)  Improved M13  phage  cloning  vectors  and  host  strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33, 103‐119.  Yoshiuchi, K., Watanabe, M. and Nishimura, A. (2000) Breeding of a non‐urea producing sake yeast with  killer character using a kar1‐1 mutant as a killer donor. J. Indust. Microbiol. Biotech. 24, 203‐209.  Zimmerli, B. and Schlatter,  J.  (1991) Ethyl  carbamate: analytical methodology, occurrence,  formation,  biological activity and risk assessment. Mutat. Res. 259, 325‐350.