Thesis (MSc)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: Glycerol metabolism is paramount to the physiological adaptation by Saccharomyces cerevisiae to hyper-osmotic stress conditions. Glycerol metabolism also -plays a fundamental role in maintaining a redox state favourable for growth under fermentative conditions. All aspects of the relationship between redox balancing and glycerol metabolism are not yet fully defined and attempts to manipulate this relationship, i.e., to increase or decrease glycerol yields from fermentation, result in a redox disturbance that is often detrimental to other aspects of metabolism. Another aspect of glycerol metabolism that is not thoroughly understood, is how the various parameters of the glycerol synthesis pathway, each independently and in conjunction with each other, control the rate at which glycerol is synthesized. Addressing these questions has been the topic of this thesis. In this regard, the theory of metabolic control analysis (MeA) was adopted and calculations were performed with the aid of an experimentally validated kinetic model. To ascertain the in vivo substrate, product, coenzyme and known modifier concentrations of the glycerol synthesis pathway, reliable techniques to halt metabolism, extract and measure these metabolites had to be established. The metabolite concentrations constitute a portion of the parameters of the pathway and are necessary to construct a detailed kinetic model. Measuring the concentration of an intracellular metabolite enzymatically requires the cell extract to have an adequate quantity of the metabolite in question. This may be achieved by concentrating the cells, before extracting the metabolite, by means of rapid filtration. Then by freezing the cells with liquid nitrogen, metabolism can be halted instantly. It was found that when metabolites were measured, yields were largely dependent on the method of extraction, since different metabolites are sensitive to different pH and temperature conditions. Methods of extraction found to be reliable for the metabolites of interest in this study are presented in Chapter 3. Metabolic control coefficients calculated by the model helped identify the parameters that control flux through the glycerol synthesis pathway most rigidly. The first reaction of the pathway, catalyzed by NAO+-dependent glycerol 3- phosphate dehydrogenase, had a flux control coefficient ( c; )of 0.83 to 0.87 and exercises the majority of control of flux through the pathway, while the subsequent reaction, catalyzed by glycerol 3-phosphatase, had far less control (C:2 = 0.13 to 0.17). The response coefficients (RJ ) of various parameter metabolites indicate [x] that flux through the pathway is most responsive to the concentration of the substrate, DHAP (RJ = 0.48 to 0.69), followed by the concentration of the [DHAP] inhibitor, ATP (RJ =-0.21 to -0.5). Interestingly, the model also predicts that [ATP] the pathway responds far more severely to the ATP/ADP ratio than to the NADH/NAD ratio, because of the weak response coefficient attributed to NADH (RJ = 0.03 to 0.08). Thus, the model suggests that the targets most strategic [NADH] for altering glycerol synthesis would be the Vmax of the glycerol 3-phosphate dehydrogenase reaction and the concentrations of DHAP and ATP. Ideally, the approach would entail manipulating each of these parameters to their optimal levels in conjunction with each other, with the least detrimental physiological effect possible. / AFRIKAANSE OPSOMMING: Gliserolmetabolisme is noodsaaklik tydens die fisiologiese aanpassing van gis onder hiperosmotiese strestoestande. Dit speelook 'n fundamentele rol in die handhawing van 'n voordelige redokstoestand, tydens groei onder fermentatiewe kondisies. Alle aspekte rondom die verwantskap tussen redoksbalansering en gliserolmetabolisme is nog nie ten volle gedefinieer nie en pogings om hierdie verwantskap te manipuleer, m.a.w. om gliserolproduksie tydens fermentasie te verhoog of verlaag, het In redoksversteuring tot gevolg, wat dikwels nadelig teenoor ander aspekte van metabolisme is. Verder is die meganisme van hoe verskeie parameters in die gliserolsintese pad, beide afsonderlik en gesamentlik, die tempo van gliserolsintese beheer, nie ten volle duidelik nie. Die doel van hierdie studie was dus om die bogenoemde onduidelikhede te probeer verklaar. In hierdie verband is die teorie van metaboliese kontrole analise (MeA) gebruik, en berekeninge is uitgevoer met behulp van 'n eksperimenteel gevalideerde kinetiese model. Ten einde, die in vivo substraat-, produk-, koensiem-, asook bekende aktiveerder en inhibeerder-konsentrasies van die gliserolsintese pad te bepaal, moes betroubare tegnieke ontwikkel word om die metabolisme vinnig te stop en sodoende metaboliete te ekstraheer en te meet. Dié metabolietkonsentrasies vorm 'n deel van die parameters in die pad, en word dus benodig om 'n gedetailleerde kinetiese model saam te stel. Die ensiematiese bepaling van intrasellulêre metabolietkonsentrasies vereis dat die selekstrak genoegsame hoeveelhede van die metaboliete bevat. Dit kan verkry word deur die selle te konsentreer deur middel van 'n vinnige filtrasiestap voordat ekstraksie van die metaboliete plaasvind. Metabolisme word onmiddellik gestop deur die selle te vries met vloeibare stikstof. Die metaboliet-hoeveelhede het grootliks afgehang van die ekstraksie-metode gebruik, aangesien verskillende metaboliete gevoelig is vir verskeie pH en temperatuurkondisies. Betroubare ekstraksiemetodes vir die metaboliete van belang vir hierdie studie word aangedui in Hoofstuk 3 van die tesis. Metaboliese kontrole koëeffisiënte wat met die model bereken is, het daardie parameters geïdentifiseer wat die fluksie deur die gliserolsintese pad die meeste beïnvloed. pie eerste reaksie in die pad, wat deur NAD+-afhanklike gliserol 3-fosfaat dehidrogenase gekataliseer word, besit 'n fluksie kontrolekoëeffisiënt ( C~) van 0.83 tot 0.87, en oefen die grootste beheer uit oor fluksie deur die pad. Die daaropvolgende reaksie, gekataliseer deur gliserol 3-fosfatase, handhaaf minder kontrole oor fluksie deur die pad (C;2 = 0.13 tot 0.17). Responskoëeffisiënte (RJ ) van verskeie parametermetaboliete dui [x) daarop dat die fluksie deur die pad die sterkste deur substraat (DHAP) konsentrasie beïnvloed word (RJ = 0.48 tot 0.69), gevolg deur inhibeerder [DHA?) (ATP) konsentrasie (RJ =-0.21 tot -0.5). Daarbenewens dui die model oak aan [AT?) dat die pad meer sensitief is teenoor die ATP/ADP verhouding, relatief tot die NADH/NAD+ verhouding, wat die gevolg is van die klein responskoëeffisiënt teenoor NADH (RJ =0.03 tot 0.08). Die model suggereer dus dat die Vmaks [NADH) van die gliserol 3-fosfaat dehidrogenase reaksie, asook die DHAP en ATP konsentrasies, die mees strategiese teikens vir manipulering van gliserolsintese behoort te wees. Die ideale benadering sal dus wees om al hierdie parameters in samehang met mekaar te kan manipuleer tot huloptimale vlakke, met die minste ontwrigting van die sel se fisiologie.

Thesis (PhD)--Stellenbosch University, 2004. / ENGLISH ABSTRACT: Key research areas in modern enology are related to the production of wine of consistent quality, as well as to the improvement of existing wine quality through the enhancement of the sensory properties of wine. The formation of glycerol during alcoholic fermentation is highly relevant to both these issues. Since the early years of the zo" century, glycerol has been positively associated with the quality of wine and it is thought to impart important mouth-feel sensations such as "viscosity", "smoothness" and "body". In general, it is considered that glycerol concentrations higher than those normally found in wine, can contribute towards the improvement of wine quality. It has also been suggested that increased concentrations of glycerol can enhance the aroma of wine. On the basis of these perceptions, several strategies have been developed to favour the production of glycerol during the fermentation process and over a period of years, a large volume of data has been collected that relates to various aspects regarding glycerol production during alcoholic fermentation. To date, however, several aspects regarding the relationship between glycerol and wine quality remain unclear. The reasons for this situation can mainly be ascribed to the lack of reliable analytical data to serve as a basis for investigating the relationship between glycerol and wine quality, as well as the preponderance of empirical and anecdotal evidence. Despite numerous opinions regarding optimal glycerol concentrations in wine, glycerol is indeed seldom assayed on a routine basis and targets with respect to specific wine grape cultivars and glycerol concentrations have largely remained unspecified. To date, very little information regarding glycerol concentrations in South African wines has been published. The analytical techniques that are most frequently used for the quantification of glycerol in grape juice, fermenting must and wine are not easily automated and this aspect placed severe limitations on the generation of large volumes of analytical data on glycerol concentrations in these matrices. This project was undertaken with the aim to holistically address some of the unresolved issues relating to the relationship between glycerol and wine quality. This also implied the development and optimisation of analytical techniques suitable for the rapid and accurate determination of glycerol in fermentation media, as well as in finished wine. In the first stage of this project a quantitative database was established that contained the analytical data on the glycerol concentrations of a statistically significant number of wines of adjudged quality, as well as additional information for each wine regarding the geographic origin, vintage, routine chemical analyses and the yeast strain(s) used for the production of the wine. The relevance of glycerol in wine for the modern South African winemaker was evaluated through the establishment of a quantitative database that contained the opinions of an expert panel of 15 South African winemakers, enologists and wine chemists on topics relating to glycerol in wine. In the second stage of the project the data captured in the databases were used to investigate aspects regarding the relationship between glycerol and wine quality. From the data captured in the qualitative database, it was clear that the topic regarding glycerol in wine was important to the South African winemakers and it was also evident that there was a need for the development and optimisation of methods suitable for the routine analysis of the glycerol concentrations in grape juice, fermenting must and wine. The opinions of the panel members also highlighted the issue that the mouth-feel property of wine was considered to be an area where the quality of some wines could be further improved. The quantitative database contained the information on the glycerol concentrations of 450 commercial South African table wines of adjudged quality. The premium cultivars Chardonnay, Chenin blanc, Sauvignon blanc, Cabernet Sauvignon, Merlot, Shiraz and Pinotage were used for the purpose of investigating the relationship between glycerol concentration and wine quality. The wines represented a wide variety of wine styles, including dry white, off-dry white, dry red and late harvest wines. The average glycerol concentration was significantly associated with the wine style. In white wines the average glycerol concentration was much lower than in the red wines (6.82 gIL versus 10.49 gIL, respectively). No significant relationship between the final glycerol concentration and the geographic origin, vintage and the yeast strain used for the fermentation was found. Wine quality could not be significantly associated with glycerol concentration in the red wines. In the white wines, the relationship between glycerol concentration and wine quality was significant, but due to the very small differences in the average glycerol concentrations of the wines of different quality ratings, the statistical significance is probably of little practical value. The effect of glycerol on the volatility of a selection of esters and higher alcohols was also investigated. Solid-phase microextraction, followed by gas chromatography, was used to analyse the composition of the headspace at equilibrium between the liquid phase and the gas phase of a model wine, and a dry white wine that contained a basal concentration of 5.4 gIL glycerol. Results showed that incremental increases in the glycerol concentrations over a range of 1 -10 gIL in the model wine and in a Chenin blanc wine, were not accompanied by a proportional increase or decrease in the abundance of the aroma components in the headspace. The volatile components tested were isoamyl acetate, ethyl butyrate, ethyl valerate, ethyl lactate, ethyl hexanoate, hexyl acetate, isoamyl alcohol and isobutanol. For all the volatile components tested, the difference between the headspace composition of samples containing the lowest glycerol concentration, and those containing the highest glycerol concentration, was not significant. However, sufficient experimental evidence was obtained to indicate that increasing glycerol concentrations had an effect on the volatility of aroma components, and that the effect is of a complex and non-linear nature. In the third phase of the work Fourier transform infrared spectroscopy (FT-IR) was used to establish and optimise methods for the accurate and rapid quantification of glycerol in wine. For this purpose calibrations were developed for the quantification of glycerol in dry wine and late harvest wines. The accuracy of prediction was evaluated by means of the standard error of prediction that was 0.38 gIL for the dry wines and 0.65 gIL for the sweet wines. Large variations are introduced in the FT-IR spectra of wine by factors such as process technology, cultivar and geographic origin, and this variation can have an effect on the accuracy of the analytical data generated when employing FT-IR spectroscopy. Using glycerol prediction in wine as a model system, principal component analysis of the FT-IR spectra was done in order to establish quality control measures for the detection of poorly predicted, or outlier samples. A classification model, based on principal component analysis, was established that enabled the interpretation and classification of the outlier samples in the data set in 100% of the cases tested. This work forms the basis for expanding the quality control measures for the detection of wines of which the FT-IR spectra are highly unnatural, as well as for establishing quality control measures to ensure that accurate analytical data are generated when FT-IR is used. FT-IR spectroscopy was also used to develop a rapid screen for the evaluation of the fermentation profiles of wine yeasts. For this purpose, a selection of wine yeasts, which included commercial wine yeasts frequently used in winemaking in South Africa, as well as a selection of hybrid Saccharomyces cerevisiae yeasts, that were obtained through a selective breeding strategy aimed at increasing glycerol concentrations were used. Calibrations necessary for the accurate quantification of glycerol, volatile acidity, ethanol, reducing sugar and glucose, in Chenin blanc must and a synthetic must were developed and optimised. This work forms the basis upon which the scope of the analysis, both in terms of the number of components that can be measured, as well as the medium in which the yeasts are being evaluated, can be enlarged. This would be valuable for future applications in both the research as well as the industrial environment. The method that was developed serves to illustrate how this application can play a supportive role in yeast development programmes, through the speeding up of the initial stages of yeast strain evaluation. / AFRIKAANSE OPSOMMING: Navorsing in moderne wynkunde is sterk gefokus op die produksie van wyn waarvan die kwaliteit van In volhoubare goeie peil is, sowel as die verdere verbetering van bestaande wynkwaliteit, deur In verhoging van die sensoriese eienskappe van wyn. Die vorming van gliserol tydens alkoholiese fermentasie het betrekking op beide hierdie aspekte. In Opvatting wat wyd gehuldig word en wat reeds sedert die vroeë jare van die 20e eeu geld, impliseer dat gliserol In positiewe bydrae lewer tot wynkwaliteit. Oor die algemeen word dit beskou dat gliserolvlakke hoër as wat normaalweg in wyn aangetref word, kan bydra om die kwaliteit van die wyn nog verder te vehoog. Daar is ook spekulasie dat verhoogde gliserolvlakke in wyn die intensiteit van die aroma van wyn kan verhoog. Hierdie opvattings het tot gevolg gehad dat veskeie strategieë ontwikkel is om die gliserol vlakke wat tydens die fermentasieproses gevorm word, te verhoog. Baie min inligting oor die gliserolinhoud van Suid-Afrikaanse wyn is tot onlangs gepubliseer. Ten spyte van die opvattings oor gliserol en wynkwaliteit wat reeds oor In lang tydperk gehuldig word, bly sekere aspekte van die verwantskap tussen gliserol en wynkwaliteit nog steeds onduidelik. Redes vir hierdie situasie kan hoofsaaklik toegeskryf word aan die totale afwesigheid van betroubare en substansiële eksperimentele data wat as basis kan dien vir die evaluering van die algemene opvattings aangaande die verwantskap tussen gliserol en wynkwaliteit. Die inligting wat wel beskikbaar is, is verder ook oorwegend van In empiriese aard. Huidig word gliserol selde op In roetine basis in die analitiese laboratorium bepaal, hoofsaaklik omdat die bestaande metodes tydrowend is en nie maklik geoutomatiseer kan word nie. In hierdie studie is In kwantitatiewe databasis opgestel waarin die inligting ten opsigte van die gliserolvlakke van 450 kommersiële Suid-Afrikaanse tafelwyne waarvan die kwaliteit beoordeel is, vervat is. Die kultivars Chardonnay, Chenin blanc, Sauvignon blanc, Cabernet Sauvignon, Merlot, Shiraz en Pinotage is gebruik vir hierdie doel en die seleksie van wyne was verteenwoordigend van In wye verskeidenheid wynstyle, insluitende droë wit-, halfdroë wit-, droë rooi- en laatoeswyne. Die gemiddelde gliserolvlakke in die witwyne was heelwat laer as die gemidelde gliserolvlakke in die droë rooiwyne (6.82 gIL teenoor 10.49 gIL, onderskeidelik). Geen beduidende verwantskap kon aangetoon word tussen die gliserolkonsentrasie in die wyn en die geografiese oorsprong, oesjaar, en die gisras wat gebruik is in die produksie van die wyn nie. Die kwaliteit van rooiwyn kon nie beduidend met die gliserolkonsentrasie geassosieer word nie. In die geval van wit wyn was die verwantskap statisties beduidend, maar die verskille was klein en moontlik nie van veel praktiese waarde nie. Die aktualiteit van gliserol vir die moderne Suid-Afrikaanse wynmaker is geëvalueer op grond van die opinies van In paneel van 15 Suid-Afrikaanse kundiges, rakende aspekte wat verband hou met die algemene opvattings oor gliserol en wyn. Die paneel het bestaan uit wynmakers, wynkundiges en chemici. Die opinie van die panel is deur middel van "n vraelys bekom en is vervat in "n kwalitatiewe databasis. Die resultate van hierdie menigspeiling het getoon dat gliserol wel vir die Suid- Afrikaanse wynmakers belangrik is en dit het ook die behoefte uitgewys vir die ontwikkeling en optimisering van metodes wat geskik is vir die roetine analyses van gliserol in wyn. Die invloed van gliserol op die vlugtigheid van "n seleksie van hoëralkohole en esters is ondersoek in "n model wyn sowel as "n wit wyn, waarvan die basiese gliserolkonsentrasie 5.4 gIL was. Soliede-fase mikroekstraksie van die gasfase van wyn is opgevolg met "n gaschromatografiese analise. Resultate het getoon dat "n stapsgewyse toename in die gliserol konsentrasie, oor "n konsentrasie reeks van 0 - 10 gIL in die model wyn, en 0 - 15 gIL in die wit wyn, nie gepaard gegaan het met "n reglynige toename of afname in die konsentrasie van die aromakomponente nie. Vir al die komponente wat ondersoek is, was die samestelling van die gasfase in monsters wat die laagste gliserol konsentrasie gehad het, teenoor dié wat die hoogste gliserol konsentrasie gehad het, nie beduidend nie. Nietemin het die resultate getoon dat gliserol wel "n effek het op die vlugtigheid van die aroma komponente wat in hierdie ondersoek gebruik is, maar dat die aard van die effek kompleks en nie-liniê is. Fourier-transformasie-infrarooispektroskopie (FT-IR) is gebruik om die metodes vir die analise van gliserol in wyn sodanig te optimiser, dat vinnige en akkurate bepalings op "n roetine basis in wyn gedoen kan word. Kalibrasies is ontwikkel vir die kwantifisering van gliserol in droëwyn en laatoeswyn. Die standaard voorspellingsfout van die kalibrasies wat ontwikkel is, was 0.38 gIL in droëwyn en 0.65 gIL in die geval van laatoeswyn. Inherente variasie in die FT-IR-spektra word deur eksterne faktore soos die proses tegnologie wat gebruik is om die wyn te berei, die kultivar en geografiese oorsprong van die wyn veroorsaak en hierdie variasie kan "n effek hê op die voorspelingsakkuraatheid van FT-IR spektroskopie. Om die effek van hierdie variasie op die akkuraarheid van die analitiese data te evalueer, is die voorspelling van gliserol as modelsisteem gebruik. Hoofkomponentanalise van die FT-IR spektra is gedoen om dié wyne met onnatuurlike FT-IR-spektra te identifiseer, ten einde beheer oor die akkuraatheid van die analitiese data uit te oefen. "n Verdere model wat ook gebaseer is op hoofkomponentanalise van die FT-IR-wynspektra is ontwikkel vir die interpretasie en klassifikasie van wynmonsters met onnatuurlike FT-IRspektra. Met die model wat ontwikkel is kon onnatuurlike wynmonsters met "n akkuraatheid van 100% gëidentifiseer word. Hierdie werk vorm die basis waarop kwaliteitsbeheer verder uitgebrei kan word sodat wyne met onnatuurlike spektra gëidentifiseer kan word, sowel as om te verseker dat akkurate data gegenereer word as FT-IR as analitiese metode gebruik word vir die kwantifisering van gliserol in wyn. FT-IR spektroskopie is ook gebruik om "n vinnige skandeerproses te ontwikkel waarmee die fermentasieprofiele van wyngiste geëvalueer kan word. Die seleksie giste wat vir hierdie doel gebruik is, sluit kommersiële Suid-Afrikaanse wyngiste in, sowel as hibriede Saccharomyces cerevisia-giste wat normaalweg hoër vlakke van gliserol tydens die fermentasieproses produseer. Kalibrasies is ontwikkel vir die akkurate kwantifisering van gliserol, vlugtige suur, alkohol, reduserende suiker en glukose, in Chenin blanc-mos sowel as in 'n sintetiese medium. Hierdie werk vorm die basis waarop verdere uitbreidings gedoen kan word, sodat meer metaboliete gemeet kan word en 'n groter verskeidenheid van fermentasiemedia gebruik kan word. Hierdie ontwikkeling is waardevol vir toekomstige toepassings in die navorsings- sowel as die industriële omgewing. Die metode wat ontwikkel is illustreer ook hoe hierdie toepassing 'n ondersteunende rol kan speel in wyngisontwikkelingsprogramme deur die aanvanklike evalueringsproses van die giste te versnel.