Quality assessment of Japanese green tea, Camellia sinensis, by techniques in metabolomics
Wipawee Pongsuwan, Bioresource engineering (Fukusaki laboratory)
Abstract
Novel techniques for Japanese green tea’s quality assessment have been developed. By utilizing techniques in metabolomics along with chemometrics, chemical information could be extracted out from complex datasets. The basic concept of this new approach is fast analysis by using chemical fingerprints instead of the characterization on a limited number of individual compounds. A number of analytical techniques were performed to explore metabolites containing in tea leaves. Hydrophilic-primary metabolites were analyzed by a coupled between gas chromatography and mass spectrometry (GC/MS). Subsequently, a system of ultra-performance liquid chromatography with mass spectrometry (UPLC/MS) was introduced to provide information of green tea’s secondary metabolites. Finally, a system of hyphenated pyrolyzer-GC/MS was developed in order to omit sample’s preparation step and explore entire metabolites comprehensively. Projection to latent structure by means of partial least-square (PLS) was performed to verify the correlation between green tea’s metabolite profiling (X matrix) and its quality (Y matrix). A quality of model was validated by testing and comparing the predictive ability to the respective model.
Chapter 1: General introduction
Quality of tea is different in kind of tea tree, plucking time, cultivation method and post harvest treatment. These variations lead to differences in metabolites content in tea leaves, thus a variation in tea’s quality. Sensory evaluation of Japanese green tea’s quality has traditionally been assessed by highly trained specialists who evaluate product quality on the basis of leave’s appearance, aroma, color and taste of the brew. Besides personal proficiency, it takes several years of specialized training to become a professional tea taster. Recently, many instrumental measurements and analysis have been introduced to evaluate and determine the quality of product by focusing on specific group compounds. However, the author believes that characteristics of product result from combination of overall metabolites rather than a single compound. Therefore, techniques in metabolomics were in the consideration.
Chapter 2: Quality-predictive model by GC/MS based primary metabolite fingerprinting
Primary metabolites of green tea, which are believed to be key factors to tea’s taste, were investigated. Green tea flavor has four characteristic taste elements: bitterness, astringency, sweetness and umami. The brothy, sweet umami taste is mainly due to amino acids. A relationship between green tea’s metabolite profile and its quality was explored. A combination of gas chromatography and mass spectrometry (GC/MS) allows the identification and quantification of numerous metabolites within single extract. A quality-predictive model based in green tea’s primary metabolite profile was constructed through PLS regression describing 82.2% of variations in Y and predicting 50.0% of variations in Y. Quinic acid, group of amino acids and sugars were found to be significant in creating the prediction model.
Chapter 3: Quality-predictive model by UPLC/MS based secondary metabolite fingerprinting
Subsequently, ultra-performance liquid chromatography coupled with mass spectrometry (UPLC/MS) was selected as an analytical platform providing information of green tea’s secondary metabolites. Several researches show that tea contains a large number of plant secondary metabolites such as catechins, purine alkaloids, flavonoids, etc. Several of them are demonstrated to be an important quality parameter determining price of product. Green tea with different quality was discriminated through principal component analysis (PCA). Consequently, projection to latent structure by means of PLS was performed. The green tea’s quality-predictive model described 83.8% of variations in Y and predicted 71.4% of variations in Y. Compounds found to have high relevance to variation in tea’s quality were group of polyphenols.
Chapter 4: Quality-predictive model of overall metabolites by PY-GC/MS
Finally, a fast, simple and low-cost approach to evaluate quality of green tea was achieved by a system of hyphenated pyrolyzer-GC/MS. The error from sample preparation could be avoided since raw samples were extracted through pyrolyzer. In addition, undesired reactions from costly derivatizing agents, which are commonly needed to treat samples before GC/MS analyses could be ignored. A quality-predictive model by means of PLS regression was created and yielded a good model with high value of goodness and fitness of the model. The model could describe 87.6% of variations in Y, while predict 77.8% of variations in Y. Differences in product’s quality were resulted from content of phenolic-derived compounds and also group of long-chain hydrocarbons.
Chapter 5: Conclusion
Techniques for quality evaluation of Japanese green tea were successfully developed by means of metabolic fingerprinting utilizing spectroscopic data along with multivariate analysis. By employing information from all metabolites containing in tea leaves, the overall image of the system was acquired.
List of publications:
This thesis is based on the following papers.
(1) Pongsuwan, W.; Fukusaki, E.; Bamba, T.; Yonetani, T.; Kobayashi, A. Prediction of Japanese green tea ranking by GC/MS based hydrophilic metabolite fingerprinting. J. Agric. Food Chem. 2007, 55, 231-236.
(2) Pongsuwan, W.; Bamba, T.; Harada, K.; Yonetani, T.; Kobayashi, A.; Fukusaki, E. A high throughput technique for comprehensive analysis of Japanese green tea quality assessment utilizing UPLC/TOFMS. (Submitted to J. Agric. Food Chem.)
(3) Pongsuwan, W.; Bamba, T.; Yonetani, T.; Kobayashi, A.; Fukusaki, E. Quality prediction of Japanese green tea using pyrolyzer coupled GC/MS based metabolic fingerprinting. J. Agric. Food Chem. 2008, 56, 744-750.