Limits...
High precision mass measurements for wine metabolomics.

Roullier-Gall C, Witting M, Gougeon RD, Schmitt-Kopplin P - Front Chem (2014)

Bottom Line: An overview of the critical steps for the non-targeted Ultra-High Performance Liquid Chromatography coupled with Quadrupole Time-of-Flight Mass Spectrometry (UPLC-Q-ToF-MS) analysis of wine chemistry is given, ranging from the study design, data preprocessing and statistical analyses, to markers identification.UPLC-Q-ToF-MS data was enhanced by the alignment of exact mass data from FTICR-MS, and marker peaks were identified using UPLC-Q-ToF-MS(2).In combination with multivariate statistical tools and the annotation of peaks with metabolites from relevant databases, this analytical process provides a fine description of the chemical complexity of wines, as exemplified in the case of red (Pinot noir) and white (Chardonnay) wines from various geographic origins in Burgundy.

View Article: PubMed Central - PubMed

Affiliation: UMR PAM Université de Bourgogne/AgroSup Dijon, Institut Universitaire de la Vigne et du Vin Jules Guyot, Dijon, France ; Research Unit Analytical BioGeoChemistry, Department of Environmental Sciences, Helmholtz Zentrum München Neuherberg, Germany.

ABSTRACT
An overview of the critical steps for the non-targeted Ultra-High Performance Liquid Chromatography coupled with Quadrupole Time-of-Flight Mass Spectrometry (UPLC-Q-ToF-MS) analysis of wine chemistry is given, ranging from the study design, data preprocessing and statistical analyses, to markers identification. UPLC-Q-ToF-MS data was enhanced by the alignment of exact mass data from FTICR-MS, and marker peaks were identified using UPLC-Q-ToF-MS(2). In combination with multivariate statistical tools and the annotation of peaks with metabolites from relevant databases, this analytical process provides a fine description of the chemical complexity of wines, as exemplified in the case of red (Pinot noir) and white (Chardonnay) wines from various geographic origins in Burgundy.

No MeSH data available.


(A) PCA score plot for the first three components from RP-UPLC-QToF-MS data sets of white wines, from three distinct appellations in Burgundy, showing the discrimination between wines according to the appellation, regardless of the age of the wine. (B) Number of metabolite structures that could be associated with elemental formulas specific to the different geographical origins, in some of the Vitis vinifera metabolic pathways, using MassTRIX.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4230200&req=5

Figure 5: (A) PCA score plot for the first three components from RP-UPLC-QToF-MS data sets of white wines, from three distinct appellations in Burgundy, showing the discrimination between wines according to the appellation, regardless of the age of the wine. (B) Number of metabolite structures that could be associated with elemental formulas specific to the different geographical origins, in some of the Vitis vinifera metabolic pathways, using MassTRIX.

Mentions: To evaluate the stability of the UPLC-Q-ToF-MS system we used QC samples and checked the RSD of peak intensity of detected peaks in all QC samples. However, we did not use this as filtering criteria because the QC samples consisted of a pool of >120 different wines from different appellations, vintages and ages and therefore the chance was high to dilute out significant and/or discriminant peaks specific to certain samples. Furthermore, samples were measured in two attempts within a timeframe of 1 year. To make both measurements comparable, we used the aligned matrix filtered for masses occurring in minimum 10% of all samples of one measurement run. In total, we obtained 977 features, among which only 692, occurring in at least 50% of all samples, were kept for further statistical analyses. However, differences in intensities between the QC samples from the two measurements were found. In order to take them into account by normalization, we used Z-transformation, which performs unit variance scaling and back transformation. This transformation strongly reduced variation across QC samples from the two measurement regimes, and thus drastically reduced, but not completely removed differences based on different times of measurement. It must be noted that the design of batches was done in a way that it resulted in sample sets that could be analyzed individually, allowing overall comparison of markers. A PCA analysis of wines was performed on RP-UPLC-Q-ToF-MS data for the four series of Burgundy Chardonnay wines, which separated them according to the appellation (Chablis, two different Meursaults and Corton Charlemagne) (Figure 5A). Although this was not the purpose of that paper, this result illustrates the potentiality of non-targeted MS-based analyses for the discrimination of wines according to their appellation of origin (geographical origin of grapes), regardless of the vintage, and thus nicely confirms that wine's metabolic diversity holds various chemical fingerprints that advanced non-targeted metabolomics can read. Possible structural assignments for specific masses for geographical origins (Figure 5B) could be obtained from databases or interfaces such as SciFinder Scholar or MassTrix (Wägele et al., 2012). However, on average more than 80% of the discriminant masses could not be annotated. Through the MassTRIX interface, 56 elemental formulas specific to Chablis wines, 59, 50, and 3 elementals formulas specific to Meursault 1, Meursault 2 and Corton Charlemagne wines, respectively (out of 304, 327, 204, and 267 masses, specific to Chablis, Meursault 1, Meursault 2, and Corton Charlemagne wines, respectively) could be correlated to hypothetical metabolite structures from the various metabolic pathways of the Vitis vinifera organism (Figure 5B) (Suhre and Schmitt-Kopplin, 2008; Wägele et al., 2012). If Chablis wines were clearly characterized by hits in each represented pathways (except for the flavanoid biosynthesis), Meursault1 wines were characterized by hits in degradation of aromatic compounds pathways and Meursault2 in Flavanoid biosynthesis pathways. All of the metabolic pathways appeared to be involved in this discrimination and covering a large number of molecular families including polyphenols, fatty acids, carbohydrates, and amino acids (Figure 5B).


High precision mass measurements for wine metabolomics.

Roullier-Gall C, Witting M, Gougeon RD, Schmitt-Kopplin P - Front Chem (2014)

(A) PCA score plot for the first three components from RP-UPLC-QToF-MS data sets of white wines, from three distinct appellations in Burgundy, showing the discrimination between wines according to the appellation, regardless of the age of the wine. (B) Number of metabolite structures that could be associated with elemental formulas specific to the different geographical origins, in some of the Vitis vinifera metabolic pathways, using MassTRIX.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4230200&req=5

Figure 5: (A) PCA score plot for the first three components from RP-UPLC-QToF-MS data sets of white wines, from three distinct appellations in Burgundy, showing the discrimination between wines according to the appellation, regardless of the age of the wine. (B) Number of metabolite structures that could be associated with elemental formulas specific to the different geographical origins, in some of the Vitis vinifera metabolic pathways, using MassTRIX.
Mentions: To evaluate the stability of the UPLC-Q-ToF-MS system we used QC samples and checked the RSD of peak intensity of detected peaks in all QC samples. However, we did not use this as filtering criteria because the QC samples consisted of a pool of >120 different wines from different appellations, vintages and ages and therefore the chance was high to dilute out significant and/or discriminant peaks specific to certain samples. Furthermore, samples were measured in two attempts within a timeframe of 1 year. To make both measurements comparable, we used the aligned matrix filtered for masses occurring in minimum 10% of all samples of one measurement run. In total, we obtained 977 features, among which only 692, occurring in at least 50% of all samples, were kept for further statistical analyses. However, differences in intensities between the QC samples from the two measurements were found. In order to take them into account by normalization, we used Z-transformation, which performs unit variance scaling and back transformation. This transformation strongly reduced variation across QC samples from the two measurement regimes, and thus drastically reduced, but not completely removed differences based on different times of measurement. It must be noted that the design of batches was done in a way that it resulted in sample sets that could be analyzed individually, allowing overall comparison of markers. A PCA analysis of wines was performed on RP-UPLC-Q-ToF-MS data for the four series of Burgundy Chardonnay wines, which separated them according to the appellation (Chablis, two different Meursaults and Corton Charlemagne) (Figure 5A). Although this was not the purpose of that paper, this result illustrates the potentiality of non-targeted MS-based analyses for the discrimination of wines according to their appellation of origin (geographical origin of grapes), regardless of the vintage, and thus nicely confirms that wine's metabolic diversity holds various chemical fingerprints that advanced non-targeted metabolomics can read. Possible structural assignments for specific masses for geographical origins (Figure 5B) could be obtained from databases or interfaces such as SciFinder Scholar or MassTrix (Wägele et al., 2012). However, on average more than 80% of the discriminant masses could not be annotated. Through the MassTRIX interface, 56 elemental formulas specific to Chablis wines, 59, 50, and 3 elementals formulas specific to Meursault 1, Meursault 2 and Corton Charlemagne wines, respectively (out of 304, 327, 204, and 267 masses, specific to Chablis, Meursault 1, Meursault 2, and Corton Charlemagne wines, respectively) could be correlated to hypothetical metabolite structures from the various metabolic pathways of the Vitis vinifera organism (Figure 5B) (Suhre and Schmitt-Kopplin, 2008; Wägele et al., 2012). If Chablis wines were clearly characterized by hits in each represented pathways (except for the flavanoid biosynthesis), Meursault1 wines were characterized by hits in degradation of aromatic compounds pathways and Meursault2 in Flavanoid biosynthesis pathways. All of the metabolic pathways appeared to be involved in this discrimination and covering a large number of molecular families including polyphenols, fatty acids, carbohydrates, and amino acids (Figure 5B).

Bottom Line: An overview of the critical steps for the non-targeted Ultra-High Performance Liquid Chromatography coupled with Quadrupole Time-of-Flight Mass Spectrometry (UPLC-Q-ToF-MS) analysis of wine chemistry is given, ranging from the study design, data preprocessing and statistical analyses, to markers identification.UPLC-Q-ToF-MS data was enhanced by the alignment of exact mass data from FTICR-MS, and marker peaks were identified using UPLC-Q-ToF-MS(2).In combination with multivariate statistical tools and the annotation of peaks with metabolites from relevant databases, this analytical process provides a fine description of the chemical complexity of wines, as exemplified in the case of red (Pinot noir) and white (Chardonnay) wines from various geographic origins in Burgundy.

View Article: PubMed Central - PubMed

Affiliation: UMR PAM Université de Bourgogne/AgroSup Dijon, Institut Universitaire de la Vigne et du Vin Jules Guyot, Dijon, France ; Research Unit Analytical BioGeoChemistry, Department of Environmental Sciences, Helmholtz Zentrum München Neuherberg, Germany.

ABSTRACT
An overview of the critical steps for the non-targeted Ultra-High Performance Liquid Chromatography coupled with Quadrupole Time-of-Flight Mass Spectrometry (UPLC-Q-ToF-MS) analysis of wine chemistry is given, ranging from the study design, data preprocessing and statistical analyses, to markers identification. UPLC-Q-ToF-MS data was enhanced by the alignment of exact mass data from FTICR-MS, and marker peaks were identified using UPLC-Q-ToF-MS(2). In combination with multivariate statistical tools and the annotation of peaks with metabolites from relevant databases, this analytical process provides a fine description of the chemical complexity of wines, as exemplified in the case of red (Pinot noir) and white (Chardonnay) wines from various geographic origins in Burgundy.

No MeSH data available.