Search

Winemaking is a biologically diverse and dynamic process in which the grape sugar is converted into ethanol, CO2 and other aromatic compounds by yeasts. Saccharomyces cerevisiae is the main species used for wine production, whereas the contribution of non-Saccharomyces yeasts to the distinctiveness of wine was not acknowledged until the 1980s. The indigenous yeasts present in the vineyard mainly belong to non-Saccharomyces species, which can have an important impact on the final wine quality, especially where spontaneous fermentation practices are used. However, metabolic profiles of individual strains of both non-Saccharomyces and Saccharomyces species may differ significantly, and thus lead to different organoleptic properties that are important to increase the expression of terroir in the wine. In this sense, some of these yeast strains may be desirable to be isolated and used for further development of novel wine products. It is also important to identify spoilage yeasts that may contaminate wine with off-flavours. Both cases require the ability to identify yeast strains that contribute particular flavour profiles to the wine. Recently, an emerging proteomic approach of matrix assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has been successfully applied to identify yeast species relevant to winemaking. This technology has shown potential in the prediction of the utility of individual yeast strains in the production of different wine styles. Despite this interest, most work focuses on its capacity for clinical identification purposes, and the list of winemaking yeasts in current MALDI-TOF databases is not exhaustive. Furthermore, the predictive potential of this approach has not gone unchallenged. With this in mind, this study aims to further develop MALDI-TOF MS as a rapid and low-cost method for yeast identification and characterisation, as well as assess it as a tool to predict the suitability of individual yeast strains in the production of different wine styles. Based on 14 type strains and 19 field isolates representing 21 yeast species, the efficiency of MALDI-TOF MS for wine yeasts identification was improved by comparing the dried-droplet (DM) and pre-mixing (PM) methods, as well as two mass ranges of m/z 2,000-20,000 and 500-4,000. With this improved protocol, MALDI-TOF MS was used to identify the yeast isolates recovered from the production of Pinot Noir wines that were spontaneously fermented in vineyard versus in winery by an organic wine producer in Waipara, New Zealand. The corresponding MALDI profiles were integrated into our in-house database stored in Software BioNumerics v 7.6. Meanwhile, 26S rRNA sequencing was used in conjunction with Restriction Fragment Length Polymorphism (RFLP) to cross-check the yeast identification results. Afterwards, eight Saccharomyces strains of diverse origin were examined to investigate the influence of growth conditions on MALDI-TOF spectra and to determine the best medium for the use of MALDI-TOF MS to predict wine yeast utility for different wine styles production, including the Pinot Noir grape juice, Chardonnay grape juice, synthetic grape juice, and laboratory-grade artificial culture media (YPD broth and agar). With the pre-selected culture media, YPD agar and YPD broth, a panel of 59 commercial yeasts including 47 wine yeasts and 12 brewing yeasts were then used to validate the predictive potential of MALDI-TOF profiling for individual yeast strains application. Dimensionality reduction techniques (DRTs) of PCA, MDS and UMAP were performed to analyse the data by using BioNumerics v 7.6 and the conda-forge packages for Python. Compared to the routine DM method, PM improved the performance of MALDI-TOF MS on wine-associated yeast analysis and yielded well-defined identification results. This is the first known usage of low-mass range m/z 500-4,000 profiles in winemaking yeast characterisation; this mass range appears unsuitable for the identification at the species level, but may offer some advantages for infraspecific (i.e. strain) classification. This improved MALDI-TOF MS protocol was then successfully applied to indigenous yeast isolated from organically produced Pinot Noir wines for diversity analysis. Thirteen species belonging to eight genera (10 non-Saccharomyces and 3 Saccharomyces yeasts) were identified, with taxonomic diversity reducing as fermentation progressed. MALDI-TOF utility also confirmed the impact of differing production systems on yeast diversity and dynamics of spontaneous fermentation. Furthermore, the MALDI profiles appeared to reflect the impact of different fermentation environments and fermentation stages on individual yeast proteomics. In addition, the yeast cultivation conditions also showed a significant impact on MALDI-TOF profiles, with YPD agar being recommended for taxonomic studies, while YPD broth may offer an improved intra-subspecific differentiation by yielding more discriminatory peaks. MDS and UMAP analyses supported the potential of MALDI-TOF proteomics in predicting the utility of yeast strains in winemaking and brewing sectors, although further studies are necessary to more comprehensively investigate the possible commercial benefits.

ABSTRACT NELSON, PAUL THOMAS. Genetic and Phenotypic Characterization of Maize Germplasm Resources: Ex-PVPA Inbreds, NCSU Inbreds, and Elite Exotic Inbreds. (Under the direction of Major M. Goodman.) Maize (Zea maize L.) germplasm resources are characterized to illuminate their usefulness and proper placement for temperate maize breeding. Three germplasm pools are examined: 1) maize inbreds that have expired U.S. plant variety protection certificates (Ex-PVPA), 2) the North Carolina State University maize inbred line releases, and 3) elite unadapted tropical maize inbreds. We have used single nucleotide polymorphism (SNP) markers to evaluate the relationships and population structure among 92 ex-PVPA inbred lines in relation to 17 well-known public inbreds. Based on UPGMA clustering, principal component analysis, and model-based clustering, we identified six primary genetic clusters represented by the prominent inbred lines B73, Mo17, PH207, A632, Oh43, and B37. We also determined the genetic background of ex-PVPA inbreds with conflicting, ambiguous, or undisclosed pedigrees. We assessed genetic diversity across subsets of ex-PVPA lines and concluded that the ex-PVPA lines are no more diverse than the public set evaluated here. The NCSU maize breeding germplasm represents a potentially useful resource for maize improvement and diversity in the U.S. While the NC maize inbreds can generally be classified into five germplasm pools, Lancaster, temperate-adapted all-tropical (TAAT), Lancaster × Tropical, Stiff Stalk, and Southern non-Stiff Stalk, analysis of detailed pedigree records and with molecular markers reveals additional substructure within each of these pools. There is general agreement among the four cluster analyses performed, three on SNP data and one on pedigree-derived coefficients of coancestry, as to the organization of this substructure. We performed topcross yield trial evaluation for 128 elite tropical maize inbreds from these institutions and 15 temperate-adapted all-tropical NC maize inbreds. We report, not only performance for yield and other traits of agronomic importance, but also heterotic patterns among many of these lines. We maintain, as reported in previous studies conducted at NCSU, that tropical germplasm, either adapted or unadapted, generally combines equally well with either Stiff Stalk or non-Stiff Stalk U.S. maize germplasm.