Introduction

Coffee is one of the world's most widely consumed beverages. To investigate the chemical characteristics of coffee, lots of analytical techniques employing gas chromatography, liquid chromatography, UV-Visible spectrophotometry, mass spectrometry, or a combination of some of those methods, have been used. By these studies, hundreds of components have been detected from coffee beans, such as caffeine, trigonelline, proteins, amino acids, carbohydrates, carboxylic acids, chlorogenic acids, lipids, glycosides and minerals. However, all the above conventional methodologies are compound-targeted, which means only one or one kind of compounds can be considered in one observation, and require appropriate extraction, purification, and chemical derivatization of each component. It is reasonable to think that even a simple treatment could cause qualitative and quantitative modifications of the original mixture. Therefore, an un-biased, non-destructive, rapid but informative method of coffee analysis is still expected to be developed.

Modern nuclear magnetic resonance (NMR) spectroscopy with its dramatically improved resolution and sensitivity has been widely applied in food chemistry to identify organic compounds non-destructively and to structurally analyze biopolymers, as it is capable in a rapid, single experiment of simultaneously detecting multiple components without destruction of the food. By identifying the different spin systems from appropriate two-dimensional (2D) NMR spectra, comprehensive analysis with NMR makes it possible to identify and quantify the components of foods as complex mixtures without physically separating them. With an assigned spectrum, the comprehensive conditions of sample can be obtained just by a single observation with sufficient qualitative, quantitative and state-related information

Therefore, the purpose of the present study is to characterize coffee bean extracts by proton (1H) and carbon ((13)C) NMR spectroscopy. Detailed signal assignments of green and roasted coffee bean extracts were carried out using various 2D NMR spectra. On the basis of signal assignments, the qualitative, quantitative and state analyses of coffee bean extracts were accomplished. Furthermore, the chemical changes over time during coffee bean roasting process, as well as the discrimination of coffee beans according to their species and origins were performed by NMR spectroscopy.

Spectral Analysis of Green Coffee Bean Extracts

The quality of coffee is directly affected by the chemical composition of green coffee beans. Therefore, a complex mixture analysis by one-dimensional (1D) and 2D NMR spectroscopy was carried out for the identification and quantification of organic compounds in green coffee bean extract (GCBE). A combination of 1H-1H DQF-COSY, 1H-(13)C HSQC, and 1H-(13)C CT-HMBC sequences was used, and 16 compounds were identified. In particular, three isomers of caffeoylquinic acid (CQA) were identified in the complex mixture without any separation. In addition, GCBE components were quantified by the integration of carbon signals by use of a relaxation reagent and an inverse-gated decoupling method without a nuclear Overhauser effect. This NMR methodology provides detailed information about the kinds and amounts of GCBE components, and in this study, the chemical makeup of GCBE was clarified by the NMR results (1).

Spectral Analysis of Roasted Coffee Bean Extracts

Roasted coffee bean extracts were characterized by 1H, (13)C NMR spectroscopy. To identify the RCBE components, a detailed and approximately 90% signal assignment was carried out using various 2D NMR spectra including 1H-1H multiple-step Relayed COSY, 1H-(13)C Edited-HSQC, 1H-(13)C CT-HMBC and a spiking method. A total of 24 coffee components, including 5 polysaccharide units, 3 stereoisomers of chlorogenic acids, and 2 stereoisomers of quinic acids, were identified with the NMR spectra of RCBE as shown in Figure 1. On the basis of the signal integration, the coffee components were successfully quantified. At the same time, state analyses were further launched for the metal ion-citrate complexes, caffeine-chlorogenate complexes and interactions between high and low molecular weight components (2).

Chemical Changes of Coffee Bean Components during Roasting Process

The roasting process of green coffee beans is probably the most important factor in the development of flavors that let us enjoy a good cup of coffee. During the roasting process, the beans undergo many complex chemical reactions which are imposable to separate, leading to important physical changes and to the formation of the substances responsible for the sensory qualities of the beverage. In this part, a 1H and (13)C NMR-based comprehensive analysis of coffee bean extracts at different degrees of roast were carried out. The roasting process of coffee bean extracts was chemically characterized using detailed signal assignment information coupled with multivariate statistical analysis. 30 NMR-visible components of coffee bean extracts were monitored simultaneously as a function of the roasting duration. As shown in Figure 2, during roasting, components such as chlorogenic acids (A) were degraded; on the other hand, components such as quinic acids (B) were generated, and some water-soluble polysaccharides (C) were generated and followed by degradation. Multivariate statistical analysis also indicated that some components, such as sucrose, chlorogenic acids, quinic acids, and polysaccharides, could serve as chemical markers during coffee bean roasting. The present composition-based quality analysis can offer an excellent holistic method and significant chemical markers to control and characterize the coffee roasting process (3).

Discrimination of Coffee Beans according to Their Species and Origins

Coffee plant belongs to the Rubiacea family, of which over 500 genera and 6,000 species exist. But of the many species, only two are of economic importance, those belonging to Coffea arabica (arabica) and Coffea cenephora, or more commonly known as robusta. Arabica accounts for roughly 75% of total world production, while robusta the remaining 25%. Therefore, in the present study, (13)C NMR spectroscopy with assigning information, coupled with principal component analysis (PCA) and orthogonal partial least square discriminant analysis (OPLS-DA) models, were employed to distinguish the species and origins of green coffee beans, and identify the significantly different metabolites between species and origins. As shown in Figure 3, green coffee bean samples were distinguished by PCA according to their species (A) and origins (B). The biomarkers such as sucrose, caffeine and chlorogenic acids were captured by OPLS-DA model (4).

Figure 1. Signal Assignments of (A)1H and (B)(13)C NMR spectra of RCBE

Figure 2. Evolusions of coffee components during roasting process.(A)chlorogenic acids;(B) quinic acids and quinids;(C)polysaccharides.

Figure 3. Discriminations of Green Coffee Beans by Their (A) Species and (B) Origins.

本論文は、NMRを用いてコーヒー豆抽出物を前処理せずにそのまま1次元および2次元NMRスペクトルを測定し、観測されたシグナルの帰属に基づいて行ったコーヒー豆成分の非破壊分析について述べている。さらに、多変量解析法と併用することによりコーヒーの化学組成の焙煎変化を解明し、コーヒー豆の品種・産地などの非破壊鑑別法を確立し、簡便かつ迅速な品質管理方法として食品産業応用に向けた展望について述べている。本論文は5章からなる。

第1章では、コーヒーとコーヒーの成分分析に関する現在までの知見について紹介し、NMRおよびNMRを食品分析に応用する利点について説明している。コーヒーは人気の高い飲料であり、質量分析などにより多種多様な成分が含まれることが報告されている。しかし、従来法では対象物の成分を分離して解析するため、実際に飲むものと異なると指摘されている。一方、NMR法は対象物の化学的特性を変化させない方法であり、成分の分離が不要で、定性、定量、状態分析が同時に行えるなどの利点から近年様々な食品の分析に利用され始めている。

第2章では、コーヒーの出発点である生豆のNMRスペクトル解析について述べられている。アラビカ種生豆抽出物の1次元の1Hと(13)C、2次元の1H-1H DQF-COSY, 1H-(13)C HSQCと1H-(13)C CT-HMBC NMRスペクトルを測定し、生豆の化学的特徴を示す生豆のNMRスペクトルを詳細に帰属することにより三種類のクロロゲン酸の異性体を含む16種類の生豆成分を定性、定量し、シグナルの緩和時間の変化から有機酸類成分の存在状態を考察している。帰属がされた生豆のスペクトルでは生豆の化学特徴を説明することが可能になった。

第3章では、中煎り程度のコーヒー焙煎豆抽出物について1次元及び2次元NMRスペクトルを測定し、焙煎豆成分の定性、定量及び状態分析について述べている。コーヒー焙煎豆抽出物のNMRスペクトルはスペクトルのパターンが極めて複雑で、帰属するために1H-1H Relayed COSYなどの様々な2次元スペクトルを測定している。その結果、クロロゲン酸の熱分解産物であるキナ酸類などの24種類の成分を帰属できた。さらに、シグナルの分離がよい(13)C NMRスペクトルを用いて定量分析を行い、帰属した焙煎豆抽出物成分の濃度を求めている。また、1H-1H ROESY測定から、クロロゲン酸とカフェインは相互作用していることを証明した。STD測定の結果、コーヒー豆抽出物に含まれている高分子物質はクロロゲン酸、カフェインなどの低分子成分と相互作用していることを確認した。

第4章では、第2章、第3章で得られた帰属情報に基づき、コーヒー生豆から焙煎度が異なるコーヒー豆の1次元NMRスペクトルを測定し、得られた各成分のシグナルの変化について述べている。シグナル積分値の変化を追跡し、得られた焙煎経時変化曲線と主成分分析法(PCA)による解析から、焙煎によって生じる化学組成の変化を解明し、帰属情報を活用することで各焙煎段階を特定できるマーカー成分を同定した。NMRスペクトルの変化から、焙煎によってクロロゲン酸類が分解し、苦味を呈するキナ酸類の成分が増加する事が確認された。また、トリゴネリンの分解と共にN-メチルプリジニウムとニコチン酸が形成される傾向が見られた。アミノ酸はショ糖と同様に急速に減少し、マンノース以外のオリゴ糖類の成分は焙煎途中で最大値に達し、焙煎がさらに進むと減少する事が確認された。帰属情報を活用することによって、NMRで観測可能な30種類のコーヒー豆成分の焙煎経時変化が明らかになった。

第5章では、NMR及びPCAを用いたコーヒー生豆の品種と産地の識別法が述べられている。コーヒー豆のNMRスペクトルには成分の種類、含量、存在状態に関する非常に多くの情報が同時に含まれていることから、NMRスペクトルとPCA解析により、2種類6産地の生豆を品種・産地ごとに区別することに成功した。また、スペクトルの帰属情報より、各品種や産地の特徴成分も特定することに成功した。この成果は、コーヒー豆の品種、原産地、偽和の有無の識別に有効な分析手法と考えられ、今後さらなる産業に向けた展望を示している。

以上、本研究はNMRを用いてコーヒー生豆および焙煎豆抽出物の特徴を示すNMRスペクトルを詳細に解析し、帰属情報を活用することでNMRによる観測可能なすべてのコーヒー豆成分の含量、存在状態、焙煎過程の経時変化および品種、産地による成分の違いの解明などに成功した。コーヒーの品質管理や他の食品の工業への応用も見据えて研究を行っており学術上、応用上貢献するところが少なくない。よって、審査委員一同は、本論文が博士(農学)の学位論文として価値あるものと認めた。