Separation and Identification of Twelve Catechins in Tea Using Liquid

Standard catechin mixtures and ... standard analytical tool for the determination of polyphe- ..... tested on a standard mixture of eight catechins an...
0 downloads 0 Views 170KB Size
Download PDF
Anal. Chem. 2000, 72, 5020-5026

Separation and Identification of Twelve Catechins in Tea Using Liquid Chromatography/Atmospheric Pressure Chemical Ionization-Mass Spectrometry Daniel J. Zeeb,† Bryant C. Nelson,‡ Klaus Albert,† and Joseph J. Dalluge*,‡

Analytical Chemistry Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, Maryland 20899-0001, and Institut fu¨r Organische Chemie, Universita¨t Tu¨bingen Auf der Morganstelle 18, 72076 Tu¨bingen, Germany

A method has been developed for the direct microscale determination of 12 catechins in green and black tea infusions. The method is based on liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry (LC/APCI-MS). Standard catechin mixtures and tea infusions were analyzed by LC/APCI-MS with detection of protonated molecular ions and characteristic fragment ions for each compound. The identities of eight major catechins and caffeine in tea were established based on LC retention times and simultaneously recorded mass spectra. In addition, monitoring of the catechin-specific retro Diels-Alder fragment ion at m/z 139 throughout the chromatogram provided a unique fingerprint for catechin content in the samples that led to the identification of four minor chemically modified catechin derivatives in the infusions. This report is the first to describe the comprehensive determination of all 12 reported catechins in a single analysis. The utility of LC/APCI-MS for providing routine separation and identification of catechins at femtomole to low-picomole levels without extraction or sample pretreatment, and its potential as a standard analytical tool for the determination of polyphenols in natural products and biological fluids, are discussed. In the last 10 years, the relationship of diet and health has become an increasingly popular subject of debate.1 While no one would argue against a role for dietary factors in the cause and prevention of important diseases including cancer and coronary heart disease, the constituents of a healthy diet have remained incompletely identified as epidemiological studies of diet and prevention of human disease are inconsonant. By contrast, identification and characterization of specific diet-derived chemicals, and an increased understanding of their biological activity, have provided support for a role of natural products and micronutrients in disease prevention and treatment. Subsequent development of these compounds as nutraceuticals will necessitate new analytical methods for their identification and measurement in * Corresponding author: NIST, 100 Bureau Drive, Stop 8392,Gaithersburg, MD 20899-8392; (fax) 301-977-0685; (e-mail) [email protected]. † Universita ¨t Tu ¨ bingen Auf der Morganstelle. ‡ National Institute of Standards and Technology. (1) Willett, W. C. Science 1994, 264, 532-537.

5020 Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

quality assessment and regulation procedures, as well as in production of appropriate standards. Direct methods for the characterization of these compounds in nature will also have a profound impact on the study of botany, microbiology, and biodiversity. Catechins are naturally occurring flavan-3-ols (Figure 1A) that are found in a variety of foods of plant origin including fruits,2 wine,3 beer,4 chocolate,5 and most abundantly tea (Camellia sinensis).6,7 The 12 reported catechins found in tea are illustrated in Figure 1B. Detailed discussions of the chemistry and health effects of tea and tea catechins are numerous12-19 and suggest a positive correlation between the dietary intake of catechins and the prevention of certain types of human disease. Methods published for the separation and identification of tea catechins include liquid chromatography (LC) combined with ultraviolet absorbance (UV), electrochemical (EC), fluorescence, and electrospray ionization mass spectrometric (ESIMS) detection, in addition to a variety of capillary electrophoretic methods.6 LC/ (2) Arts, I. C. W.; Hollman, P. C. H. J. Agric. Food Chem. 1998, 46, 51565162. (3) Carando, S.; Teissedre, P. L.; Pascual-Martinez, L.; Cabanis, J. C. J. Agric. Food Chem. 1999, 47, 4161-4166. (4) Madigan, D.; McMurrough, I.; Smyth, M. R. Analyst 1994, 119, 863-868. (5) Arts, I. C. W.; Hollman, P. C. H.; Kromhout, D. Lancet 1999, 354, 488488. (6) Dalluge, J. J.; Nelson, B. C. J. Chromatogr., A 2000, 881, 411-424. (7) Dalluge, J. J.; Nelson, B. C.; Thomas, J. B.; Sander, L. C. J. Chromatogr., A 1998, 793, 265-274. (8) Registry File Database, Chemical Abstracts Service; American Chemical Society: Washington, DC, 1999. (9) Davis, A. L.; Cai, Y.; Davies, A. P.; Lewis, J. R. Magn. Reson. Chem. 1996, 34, 887-890. (10) Miketova, P.; Schram, K. H.; Whitney, J.; Li, M.; Huang, R.; Kerns, E.; Valcic, S.; Timmermann, B. N.; Rourick, R.; Klohr, S. J. Mass Spectrom. 2000, 35, 860-869. (11) Saijo, R. Agric. Biol. Chem. 1982, 46, 1969-1970. (12) Stavric, B. Clin. Biochem. 1994, 27, 319. (13) Wiseman, S. A.; Balentine, D. A.; Frei, B. Crit. Rev. Food Sci. Nutr. 1997, 37, 705. (14) Weisburger, J. H. Proc. Soc. Exp. Biol. Med. 1999, 220, 271. (15) Hollman, P. C.; Feskens, E. J.; Katan, M. B. Proc. Soc. Exp. Biol. Med. 1999, 220, 198. (16) Yang, C. S.; Wang, Z. Y. J. Natl. Cancer Inst. 1993, 85, 1038. (17) Balentine, D. A. Manufacturing and chemistry of tea. In Phenolic Compounds in Food and their Effects on Health I: Analysis, Occurrence and Chemistry; Ho, C. T., Lee, C. Y., Huang, M. T., Eds.; American Chemical Society: Washington, DC, 1992; p 102. (18) Ahmad, N.; Mukhtar, H. Nutr. Rev. 1999, 57, 78. (19) Fujiki, H. J. Cancer Res. Clin. Oncol. 1999, 125, 589-597. 10.1021/ac000418f CCC: $19.00

© 2000 American Chemical Society Published on Web 09/06/2000

Figure 1. Catechin structures. (A) General structure of a catechin. (B) Structures of the 12 tea catechins identified in this study.

ESIMS has been the method of choice for the mass spectrometric determination of catechins in complex mixtures,20,21 while fast atom bombardment (FAB)-, electron impact (EI)-, and ESI-MS/ MS have all been employed to provide both molecular mass and structural information for these analytes.21,22 Interestingly, there is no extant literature on methods for the comprehensive separation and determination of all 12 tea catechins. This is probably due at least in part to the paucity of published data regarding the (20) Dalluge, J. J.; Nelson, B. C.; Thomas, J. B.; Welch, M. J.; Sander, L. C. Rapid Commun. Mass Spectrom. 1997, 11, 1753. (21) Miketova, P.; Schram, K. H.; Whitney, J. L.; Kerns, E. H.; Valcic, S.; Timmermann, B. N.; Volk, K. J. J. Nat. Prod. 1998, 61, 63. (22) Poon, G. K. J. Chromatogr., A 1998, 794, 63-74.

distribution of minor catechin components (chemically modified EC, ECG, and EGCG derivatives) in tea, as well as to a lack of sensitivity and selectivity of previously published methods. Recent success with the use of directly combined liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry (LC/APCI-MS) for the determination of flavanones and xanthones in crude plant extracts23 suggests that this technique might be equally effective in the analysis of tea polyphenols. The present report details the use of LC/APCI-MS for the on-line determination of the 12 known catechins in green (23) da Costa, C. T.; Dalluge, J. J.; Welch, M. J.; Coxon, B.; Margolis, S. A.; Horton, D. J. Mass Spectrom. 2000, 35, 540-549.

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5021

and black tea infusions. The method demonstrates the efficient and selective separation and detection of analytes characteristic of LC/MS analysis, while providing at least 2 orders of magnitude higher sensitivity than analogous LC/UV6,7,24 and LC/ESIMS methods20,21 for the identification of polyphenols in tea. In addition, it is demonstrated that fragmentation of APCI-generated ions by application of an electrostatic potential at the entrance to the single quadrupole mass analyzer produces a fragment ion specific to catechins. Monitoring of this ion throughout the chromatogram permits detection of both major (eight) and minor (four) catechins present in a complex mixture. Subsequent mass spectral analysis yields molecular mass and structural information for each of the analytes. EXPERIMENTAL SECTION Note: Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the material or equipment identified are the best available for the purpose. Chemicals. (-)-Epigallocatechin (EGC), (-)-gallocatechin (GC), (+)-catechin (C), (-)-catechin gallate (CG), caffeine, (-)epigallocatechin gallate (EGCG), (-)-epicatechin (EC), (-)gallocatechin gallate (GCG), (-)-epicatechin gallate (ECG), and trifluoroacetic acid (TFA) were purchased from Sigma Chemical Co. (St. Louis, MO). HPLC-grade methanol (MeOH) and acetonitrile (ACN) were purchased from J. T. Baker (Baker-Mallinckrodt, Phillipsburg, NJ). HPLC-grade water (18 mΩ), prepared using a Millipore Milli-Q purification system (Millipore Corp., Bedford, MA), was used to prepare all solutions. Preparation of Catechin Standards. Standard stock solutions of each catechin and caffeine (3 mg/mL), and a solution containing 0.05 µg/µL of each of the eight catechins and caffeine in a 30% volume fraction of methanol were prepared and used for methods development. Preparation of Tea Samples. Six commercially available teas were examined for catechin content employing LC/APCI-MS. Green teas investigated included a Darjeeling green tea (Whittard of Chelsea, London, England), and the China green teas Lung Ching, Pi Lo Chun, and Snow Dragon (SpecialTeas, Inc., Norwalk, CT). Black teas investigated included a Darjeeling blend and a Sonarie Assam. For each variety of tea, ∼200 mg of tea leaves was steeped at 80 °C for 10 min in 20 mL of water. After cooling for 5 min, samples were filtered through a 0.2-µm nylon filter and analyzed directly by LC/APCI-MS. Liquid Chromatography/Atmospheric Pressure Chemical Ionization-Mass Spectrometry. Analyses of the standard catechin mixtures and tea extracts were carried out using a HewlettPackard Series 1100 LC/MSD instrument (Wilmington, DE) consisting of an HP 1100 series liquid chromatograph with a variable-wavelength UV absorbance monitor placed in series between the chromatograph and the single quadrupole mass spectrometer equipped with an APCI source. LC separations were made according to the method of Dalluge et al. 7 using a 4.6 × 250 mm Zorbax Eclipse XDB-C18 reversed-phase chromatography column (Mac-Mod Analytical, Chadds Ford, PA) at room temper(24) Goto, T.; Yoshida, Y.; Kiso, M.; Nagashima, H. J. Chromatogr., A 1996, 749, 295-299.

5022

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

ature. The LC mobile phase consisted of (A) water containing a volume fraction of 0.05% TFA and (B) ACN containing a volume fraction of 0.05% TFA. The gradient was linear from 12 to 21% B, 0-25 min, linear from 21 to 25% B, 25-30 min, and linear from 25 to 100% B, 30-35 min. The injection volume was 10 µL for the standard solutions and 20 µL for the tea extracts. The flow rate was 1 mL/min, and UV absorbance was monitored at 280 nm. Several parameters of the APCI interface were optimized by flow injection analysis (FIA) of standard catechin solutions. The sensitivity of detection of catechins in positive ion APCI mode was found to be at least 1 order of magnitude higher than for negative ion APCI. Therefore, instrumental parameters were selected that maximized generation of the protonated molecular ion [M + H]+ of each analyte and that also produced characteristic fragment ions. Parameters optimized included the corona needle current, the capillary voltage, and the fragmentor potential. The following instrumental parameters were used for APCI-MS detection of the catechins in the positive ion mode: corona needle, 10 µA; capillary voltage, 4000 V; fragmentor, 60 V; nebulizer pressure, 414 kPa (60 psi); gain, 3; threshold, 100; step size, 0.15; drying gas, 3.5 mL/min; drying gas temperature, 350 °C; vaporizer temperature, 500 °C. Uncertainties for reported mass/charge ratios and molecular masses are (0.01%. For comparison of APCIMS detection to existing ESI-MS-based methods, flow injection analyses of the catechins were performed in positive and negative ESI-MS modes as previously described.20-22 RESULTS AND DISCUSSION Current interest in the investigation of tea catechins as a source of new pharmaceutical agents has stemmed from compelling evidence of the antioxidant,13,25 anticarcinogenic,16 and antimicrobial26,27 activities of these compounds. As catechins begin undergoing preclinical efficacy and toxicity testing, analytical approaches that address the comprehensive determination of these 12 polyphenols (Figure 1) in complex mixtures must be developed. Previous studies have demonstrated the potential of LC/APCIMS for the identification of flavonoids in plant extracts.23 Examination of the general utility of this technique for high-sensitivity determination of catechins and identification of minor chemically modified catechin derivatives in tea infusions was the specific aim of this study. Determination of Catechins in a Standard Mixture Using LC/APCI-MS. Prior to examining LC/APCI-MS for the determination of catechins in crude tea infusions, the efficacy of the separation and detection of catechins using this technique was tested on a standard mixture of eight catechins and caffeine. The results of this experiment are illustrated in Figure 2. Figure 2A shows a total ion chromatogram for this analysis that demonstrates baseline separation of the nine-component mixture within 25 min utilizing a previously described LC elution system.7 The identities of the nine components were established from LC retention times and from continuously recorded mass spectra. Alignment of characteristic mass spectral ion profiles (Figure 2B-F) with UV (25) Salah, N.; Miller, M. J.; Paganga, G.; Tijburg, L.; Bolwell, G. P.; Rice-Evans, C. Arch. Biochem. Biophys. 1995, 322, 339-346. (26) Sakanaka, S,; Kim, M.; Taniguchi, M.; Yamamoto, T. Agric. Biol. Chem. 1989, 53, 2307-2311. (27) Nakayama, M.; Suzuki, K.; Toda, M.; Okubo, S.; Hara, Y.; Shimamura, T. Antiviral Res. 1993, 21, 289-299.

Figure 2. LC/APCI-MS for the separation and identification of eight catechins and caffeine in a standard mixture. The concentration of each catechin was 0.05 µg/µL; 10-µL injection. (A) Total ion chromatogram. Peak identification: (1) GC; (2) EGC; (3) caffeine; (4) C; (5) EC; (6) EGCG; (7) GCG; (8) ECG; (9) CG. (B) Reconstructed ion chromatogram for m/z 291 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). (C) Reconstructed ion chromatogram for m/z 307. (D) Reconstructed ion chromatogram for m/z 443. (E) Reconstructed ion chromatogram for m/z 459 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). (F) Reconstructed ion chromatogram illustrating the catechin-specific RDA fragment ion at m/z 139.

detection peak profiles (not shown) allows unambiguous peak assignment for each analyte in the standard mixture. Characteristic mass spectra for (+)-catechin and (-)-epigallocatechin gallate are illustrated as insets to panels B and E of Figure 2, respectively. The mass spectrum of C (Figure 2B, inset) includes the protonated molecular ion [M + H]+ at m/z 291, as well as characteristic fragment ions at m/z 273 [M + H - H2O]+ and 139 (see below). This fragmentation pattern is consistent for all the catechin and gallocatechin species measured. The mass spectrum for EGCG (Figure 2E, inset) shows the detection of [M + H]+ at m/z 459 and characteristic fragment ions at m/z 289 [M + H - galloyl + H - H2O]+ and 139 (see below). This general fragmentation pattern was observed for all of the catechin gallates and gallocatechin gallates measured in this study. In general, higher fragmentation voltages (higher electrostatic potentials at the entrance to the mass analyzer) led to an increase in characteristic fragment ion intensities that was proportional to the decrease in intensity of the corresponding protonated molecular ion for all analytes tested. Figure 2F illustrates a reconstructed ion chromatogram for m/z 139 in the analysis of the standard mixture. This ion arises from a retro Diels-Alder (RDA) fragmentation of the nonvariable portion of the catechin ring structure. The pathway for this fragmentation, previously described for the characterization of catechins using EI-MS21 and FAB-MS28 is shown in Figure 3. Production of this fragment ion in the mass analysis of catechins using the atmospheric pressure ionization techniques of APCI(28) Stobneicki, M.; Popenda, M. Phytochemistry 1994, 37, 1707-1711.

MS or ESI-MS, however, has not been described prior to this report. In support of the proposed fragmentation mechanism, alignment of the reconstructed ion chromatogram illustrated in Figure 2F with the total ion chromatogram in Figure 2A indicates that m/z 139 is common among all of the catechin species determined. Monitoring of this ion throughout a chromatogram should therefore provide an immediate fingerprint for identification of catechins present in complex mixtures. A demonstration of the sensitivity of LC/APCI-MS in the measurement of catechins is illustrated in Figure 4. In this experiment, 15 pmol (=5 ng, 10-µL injection) of C and EC was separated and selectively detected in a dilute standard catechin mixture. Although detection of smaller amounts of material was not attempted, the strong signal illustrated in Figure 4 suggests limits of detection for this method in the femtomole to lowpicomole range. Analysis of similar catechin preparations using previously reported positive and negative ion ESI-based methods20-22 was also performed (not shown). A comparison of the data indicated that APCI-MS is at least 2 orders of magnitude more sensitive than ESI-MS for the detection of catechins, based on overall signal intensity and signal-to-noise ratios. This level of sensitivity combined with the selectivity of mass spectrometric detection will certainly facilitate routine measurement of catechins in biological fluids using LC/APCI-MS, where concentrations after ingestion of tea range from 50 to 250 ng/mL.29 Research in this area is currently underway in our laboratory. Determination of Catechins in Tea Infusions Using LC/ APCI-MS. The utility of the developed LC/APCI-MS system for separation and detection of catechins in commercially available green and black teas was also tested. Sample preparation simulated actual brewing conditions for a cup of tea, and filtered infusions were analyzed directly without further pretreatment or extraction. The total ion chromatogram for the analysis of the China green tea Lung Ching is shown in Figure 5A. Identities of the eight major tea catechins and caffeine in each infusion were established from LC retention times relative to the standard mixture (Figure 2) and from continuously recorded mass spectra. Reconstructed ion chromatograms and characteristic mass spectra for the catechins EGCG and ECG are shown in panels B and C of Figure 5, respectively. Comparison of the reconstructed ion chromatogram and corresponding mass spectrum of EGCG shown in Figure 5B (green tea infusion) with that shown in Figure 2E (standard mixture) attests to the accuracy with which peak assignments can be made using this method. The eight major catechins (C, EC, GC, EGC, CG, ECG, GCG, EGCG) were all present in five of the six teas tested. The exception was the Sonarie Assam black tea that contained very low levels of six catechins and undetectable levels of GCG and CG (Table 1). This is probably the result of the extensive fermentation required for production of this tea that is known to cause significant oxidative degradation of catechins.16 The utility of the reconstructed ion chromatogram of m/z 139 as a fingerprint for catechin content in complex mixtures is obvious from Figure 5D. Close inspection of this chromatogram suggested the presence of several minor catechin components in the Lung Ching tea extract at approximate retention times of 17.9, (29) Lee, M. J.; Wang, Z. Y.; Li, H.; Chen, L.; Sun, Y.; Gobbo, S.; Balentine, D. A.; Yang, C. S. Cancer Epidemiol. Biol. Prev. 1995, 4, 393-399. (30) Tuomi, T.; Johnsson, T.; Reijula, K. Clin. Chem. 1999, 45, 2164-2172.

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5023

Figure 3. Retro Diels-Alder fragmentation of a catechin.

Figure 4. Sensitivity of LC/APCI-MS for the determination of catechins demonstrated using selected ion monitoring of m/z 291 for detection of C and EC. Ten-microliter injection of a standard mixture containing 0.5 ng/µL of eight catechins and caffeine. Table 1. Identification of 12 Catechins in 6 Commercially Available Loose Leaf Teasa green tea

black tea

Lung Pi Lo Snow Darjeeling Darjeeling Sonarie Ching Chun Dragon green black Assam C EC GC EGC CG ECG GCG EGCG EZ EZG ECMG EGCMG

+ + + + + + + + + + + +

+ + + + + + + + + + -

+ + + + + + + + + + -

+ + + + + + + + + + + +

+ + + + + + + + + + + +

+ + + + + + + + + -

a +, compound present in corresponding tea; -, compound absent from corresponding tea. Quantification of catechins in these teas was not performed.

19.8, 30.5, and 31.3 min. Selected ion profiles and corresponding mass spectra of the modified catechins eluting at these positions are shown in Figure 6A-D. The average mass spectrum recorded under the peak eluting at 17.9 min is shown in the inset to Figure 6A. The molecular mass of this component, 274 g/mol, suggested a loss of oxygen from C or EC and was tentatively assigned as the previously reported catechin derivative epiafzelechin (EZ).8,9 This assignment is corroborated by comparison of the fragmentation pattern of EZ (Figure 6A, inset) with that of EC (Figure 2B, inset), the LC retention time of this compound relative to EC, and the presence of the catechin-specific fragment ion at m/z 139. Designation of the epimer configuration of this and the modified 5024

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

catechins discussed below as “epi” is based on the relative abundance of the “epi” configuration in known polyphenolic components of tea infusions and extracts.7,24 Mass spectral analysis of the peak eluting at 31.3 min identifies a protonated molecular ion at m/z 427 (Figure 6B). The fragmentation pattern for this compound (Figure 6B, inset) shows the catechin-specific RDA fragment ion at m/z 139 and the [M + H - galloyl + H - H2O]+ ion, characteristic of catechin gallates, at m/z 257. This, in addition to an LC retention time relative to ECG that is consistent with previous reports on the identification of this catechin in tea extracts,10 and a mass shift of 152 Da (galloyl ) 152 Da) relative to EZ (Figure 6A), allows identification of this analyte as epiafzelechin gallate (EZG). The question arose as to which oxygen of ECG is not present in the structure of EZG. The presence of the [M + H - galloyl + H - H2O]+ ions ([M + H - 170]) for both ECG and EZG demonstrates that the oxygen is not removed from the gallic acid moiety. Previous investigations of EZG in extracts from kilogram quantities of tea using LC/ESIMS have been unable to distinguish whether the oxygen was missing from the A or B ring of EZG10 (Figure 1A). By contrast, the unaltered presence of the RDA fragment at m/z 139 in the mass spectrum of EZG illustrated in Figure 6B of this report supports the absence of an oxygen on the B ring, consistent with what is known regarding the biosynthesis of catechins.31 Analysis of the mass spectral data for each of the peaks in the m/z 139 reconstructed ion chromatogram also revealed the presence of a component eluting at 19.8 min with a [M + H]+ ion at m/z 473. The reconstructed ion chromatogram for m/z 473 and the average mass spectrum corresponding to this peak are illustrated in Figure 6C. The mass shift of 14 Da relative to EGCG led to the tentative assignment of this compound as the previously reported (-)epigallocatechin methylgallate (EGCMG).9-11 This assignment is supported by the following observations. First, the presence of m/z 139 establishes that the compound is structurally related to a catechin and that the A ring is unmodified. Second, the presence of m/z 289 establishes the relationship of this compound to EGCG (compare Figure 6C, inset to Figure 5B, inset) and indicates that the A and B rings are both unmodified, leaving the gallic acid moiety as the site of methylation. Previous studies have demonstrated that an unmodified 4′′-hydroxyl (Figure 1A) is required for the loss of gallic acid from catechin gallates to form m/z 289.10 NMR studies of catechins extracted from tea show that methylation of EGCG is at the 3′′-hydroxyl group.9 Definitive assignment of the gallate-methylated EGCG identified (31) Mann, J. Secondary Metabolites; Claredon Press: Oxford, 1987.

Figure 5. LC/APCI-MS analysis of the China green tea Lung Ching. (A) Total ion chromatogram. Peak identification: (1) GC; (2) EGC; (3) caffeine; (4) C; (5) EC; (6) EGCG; (7) GCG; (8) EZ; (9) EGCMG; (10) ECG; (11) CG; (12) ECMG; (13) EZG. (B) Reconstructed ion chromatogram for m/z 459, and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). (C) Reconstructed ion chromatogram for m/z 443 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). (D) Reconstructed ion chromatogram for m/z 139. Peak identification as in (A). The m/z 139 peak eluting at 8.3 min (3) arises from the first isotope of the [M + H - C2NOH3]+ fragment ion of caffeine.30

Figure 6. Identification of four modified catechins in the China green tea Lung Ching using LC/APCI-MS. (A) Reconstructed ion chromatogram for m/z 275 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). The m/z 275 peak eluting at 17.9 min corresponds to the protonated molecular ion of EZ. The m/z 275 peak eluting at 23.8 min corresponds to the second isotope peak of the [M + H - 170]+ ion originating from ECG. (B) Reconstructed ion chromatogram for m/z 427 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). Unlabeled peaks were shown from their corresponding mass spectra not to be catechins. (C) Reconstructed ion chromatogram for m/z 473 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). (D) Reconstructed ion chromatogram for m/z 457 and simultaneously recorded mass spectrum averaged across the peak labeled with an asterisk (inset). Unlabeled peaks were shown from their corresponding mass spectra not to be catechins.

here as (-)-epigallocatechin 3′′-O-methylgallate, however, is unwarranted. The average mass spectrum recorded across the peak eluting at 30.5 min (Figure 6D) identifies the analogous monomethylated derivative of ECG, (-)-epicatechin methylgallate (ECMG). This assignment is corroborated by LC retention relative to ECG, mass spectral data consistent with a derivative of ECG that has been monomethylated on the gallic acid moiety (compare Figures 5C and 6D), and the presence of the catechin-specific m/z 139 RDA fragment ion. The presence of the eight major catechins and the four minor chemically modified catechin derivatives

described above in six commercially available green and black teas is summarized in Table 1. These results represent the first comprehensive determination of the 12 known catechins in a single analysis using any method. In addition, the highly sensitive LC/APCI-MS method described herein can selectively identify the eight major and four minor catechins directly in tea infusions. This achievement could not have been realized using less selective LC/UV, LC-EC, or LCfluorescence methods and has yet to be accomplished using LC/ ESI-MS. Previous reports describing characterization of the four Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

5025

low-abundance chemically modified catechin derivatives EZ, EZG, EGCMG, and ECMG using the less sensitive techniques of LC/ ESI-MS and NMR required analysis of polyphenolic extracts from kilogram quantities of green tea.9,10 By contrast, the current method is shown to be applicable to the routine measurement of tea catechins in the femtomole to low-picomole range without extraction or sample pretreatment. CONCLUSIONS The LC/APCI-MS method described in this report should find widespread use for detection of catechins in tea and other foods of plant origin. The sensitivity of the technique, together with the use of an appropriate internal standard, will be the basis of a method for the accurate quantification of catechins in these matrixes, as well as in biological fluids. The application of LC/ APCI-MS to the measurement of catechins in blood plasma should be useful to clinical researchers investigating the bioavailability

5026

Analytical Chemistry, Vol. 72, No. 20, October 15, 2000

and pharmacokinetics of these flavonoids, as well as links between tea consumption and disease prevention or treatment. Available LC/APCI-MS instrumentation is small, affordable, and simple to use. The technique described here is therefore immediately accessible not only to the mass spectrometrist or separations scientist but also to anyone interested in the rapid identification and characterization of flavonoids in complex mixtures. Finally, the utility of this technique, as described here and in previous reports23 for the determination of a wide variety of polyphenols, should prove useful for regulation and quality control monitoring of nutraceuticals, as well as in the identification of the biologically active constituents of natural products and the diet.

Received for review April 10, 2000. Accepted July 30, 2000. AC000418F