Systematic metabolite annotation and identification in complex biological extracts : combining robust mass spectrometry fragmentation and nuclear magnetic resonance spectroscopy
Hooft, J.J.J. van der - \ 2012
Wageningen University. Promotor(en): Raoul Bino; Sacco de Vries, co-promotor(en): Jacques Vervoort; Ric de Vos. - S.l. : s.n. - ISBN 9789461732347 - 256
metabolieten - metabolomica - massaspectrometrie - kernmagnetische resonantiespectroscopie - metabolische profilering - metabolische fingerprinting - metabolites - metabolomics - mass spectrometry - nuclear magnetic resonance spectroscopy - metabolic profiling - metabolic fingerprinting
Detailed knowledge of the chemical content of organisms, organs, tissues, and cells is needed to fully characterize complex biological systems. The high chemical variety of compounds present in biological systems is illustrated by the presence of a large variety of compounds, ranging from apolar lipids, semi-polar phenolic conjugates, toward polar sugars. A molecules’ chemical structure forms the basis to understand its biological function. The chemical identification process of small molecules (i.e., metabolites) is still one of the major focus points in metabolomics research. Actually, no single analytical platform exists that can measure and identify all existing metabolites. In this thesis, two analytical techniques that are widely used within metabolite identification studies have been combined, i.e. mass spectrometry (MS) and nuclear magnetic resonance spectroscopy (NMR). MS was used to ionize the metabolites and to record their molecular weight and to provide substructure information based on fragmentation in the mass spectrometer. NMR gave the comprehensive structural information on the chemical environment of protons and their linkage to other protons within the molecule. The additional structural information as compared to MS is at the cost of an increased amount of compound needed for NMR detection and spectra generation. Here we combined both analytical methods into a liquid chromatography (LC)-based platform that concentrated compounds based on their specific mass; thereby providing a direct link between MS and NMR data. Another platform was developed that generated robust multistage MSn data, i.e., the systematic fragmentation of metabolites and subsequent fragmentation of resulting fragments.
This thesis aims to accelerate metabolite identification of low abundant plant and human derived compounds by following a systematic approach. The acquired structural information from MSn and 1D-1H-NMR spectra resulted in the complete elucidation of phenolic metabolites in microgram scale from both plant and human origin.
In the chapter 1, the analytical techniques and terms used throughout the thesis are introduced. The second chapterdescribes how a high mass resolution MSn fragmentation approach was tested in both negative and positive ionization modes for differentiation and identification of metabolites, using a series of 121 polyphenolic molecules. An injection robot was used to infuse the reference compounds one by one into a hybrid mass spectrometer, combining MSn possibilities with accurate mass read-out. This approach resulted in reproducible and robust MSn fragmentation trees up to MS5, which were differential even for closely related compounds. Accurate MSn-based spectral trees were shown to be robust and powerful to distinguish metabolites with similar elemental formula (i.e. isomers), thereby assisting compound identification and annotation in complex biological samples. In the third chapter, we tested the annotation power of this spectral tree approach for annotation of phenolic compounds in crude extracts from Lycopersicum esculentum(tomato) and the model plant Arabipopsis thaliana. Partial MSn spectral trees were generated directly after chromatographic elution (LC-MSn). Detailed MSn spectral trees could be recorded with the use of a collector/injector robot.We were able to discriminate flavonoid glycosides based on their unique MSn fragmentation patterns in either negative or positive ionization mode. Following this approach, we could annotate 127 metabolites in the tomato and Arabidopsis extracts, including 21 novel metabolites. The good quality MSn spectral trees obtained can be used to populate MSn databases and the protocols to generate the spectral trees are a good basis to further expand this database with more diverse compounds.
Chapter 4 then describes how an automated platform, coupling chromatography with MS and NMR (LC-MS-solid phase extraction-NMR), was developed that can trap and transfer metabolites based on their mass values from a complex biological extract in order to obtain NMR spectra of the trapped LC-MS peak, out of minute amounts of sample and analyte. Extracts from tomatoes modified in their flavonoid biosynthesis pathway were used as proof of principle for the metabolite identification process. This approach resulted in the complete structural elucidation of 10 flavonoid glycosides. This study shows that improving the link between the mass signals and NMR peaks derived from the selected LC-MS peaks decreases the time needed for elucidation of the metabolite structures. In addition, automated 1D-1H-NMR spectrum fitting of the experimental data obtained in this study using the PERCH NMR software further speeded up the candidate rejection process.
Chapter 5 illustrates how the two developed analytical platforms could be used for the successful selection, annotation, and identification of 177 phenolic compounds present in different extracts of Camellia sinensis, i.e. green, white, and black tea extracts, including the full identification of microgram amounts of complex acylated conjugates of kaempferol and quercetin. Principal component analysis based on the relative abundance of the annotated phenolic compounds in 17 commercially available black, green and white tea products separated the black teas from the green and white teas, thereby illustrating the differential phenolic metabolite contents of black tea as compared to green and white teas. The change in phenolic profiles reflects the polymerization reactions occurring upon transformation of green tea into black tea. This study shows that the combined use of MSn spectral trees and LC-MS-solid phase extraction-NMR leads to a more comprehensive metabolite description thereby facilitating the comparison of tea and other plant samples.
In chapter 6, we aimed to structurally elucidate and quantify polyphenol-derived conjugates present in the human body by studying the urinary excretion of these conjugates.We applied a combination of a solid phase extraction preparation step and the two HPLC-coupled analytical platforms as described in chapters 2 and 3. This analytical strategy resulted in the annotation of 138 urinary metabolites including 35 completely identified valerolactone conjugates. These valerolactones are microbial break-down products of tea phenols. NMR predictions of glucuronidated and sulphonated core metabolites were performed in order to confirm the NMR peak assignments on the basis of 1D-1H-NMR data only. In addition, 26 hours quantitative excretion profiles for certain valerolactone conjugates were obtained using diagnostic proton signals in the 1D-1H-NMR spectra of urine fractions.
In the seventh chapter, the current state of metabolite identification and expected challenges in the structural elucidation of metabolites at (sub)microgram amounts are discussed. The work in this thesis and of other groups working on the hyphenation of MS and NMR shows that the complete de novo identification of microgram amounts and even lower of compound is feasible by using MS guided solid phase extractiontrapping in combination with 1D-1H-NMR or UPLC-TOF-MS isolation followed by capillary NMR. Semi-automated annotation of compounds based on their MS and NMR features is now feasible for some well studied compound classes and groups.
Altogether, the developed platforms yield new and improved insights in the phenolic profiles of well-studied plants as well as a comprehensive picture of the metabolic fate of green tea polyphenols upon intake in the human body. The followed metabolite identification strategy is useful for other studies that aim to elucidate bioactive compounds, especially when only small sample volumes are available. This thesis also contributes to the acquisition of good quality data for metabolite identification by acquiring robust MSn fragmentation spectra and 1D-1H-NMR spectra of partial purified analytes at microgram scale, which paves the path for further developments in data acquisition and analysis, as well as the unravelling of yet unknown metabolites in a faster, more systematic and automated manner.