Abstract
Accurate measurements of the secondary metabolites in natural products and plant foods are critical for establishing relations between diet and health. There are as many as 50,000 secondary metabolites that may influence human health. Their structural and chemical diversity presents a challenge to analytical chemistry. With respect to flavonoids, putative identification is accessible, but positive identification and quantification are limited by the lack of standards. Quantification has been tested with use of both nonspecific and specific methods. Nonspecific methods, which include antioxidant capacity methods, fail to provide information on the measured components, suffer from numerous interferences, are not equatable, and are unsuitable for health research. Specific methods, such as LC with diode array and mass spectrometric detection, require the use of internal standards and relative molar response factors. These methods are relatively expensive and require a high level of expertise and experimental verification; however, they represent the only suitable means of relating health outcomes to specific dietary components.
Keywords: dietary flavonoids, antioxidants, mass spectrometry, proanthocyanidins, secondary metabolites
Introduction
Accurate measurement, defined as identification and quantification, is particularly critical for natural products. Unlike commonly consumed fruits and vegetables, the potentially health-promoting components of interest in natural products are not the vitamins and minerals but the large variety of secondary metabolites. That is not to say that the secondary metabolites of fruits and vegetables are not important. Steinmetz and Potter (1) reported that a diet rich in fruits and vegetables correlated with a reduced risk of chronic diseases. This correlation was not as strong for traditional vitamins and has given rise to an increased interest in secondary metabolites. Unlike vitamins, the impact of plant secondary metabolites on health may require years to become apparent (2). Thus, accurately analyzing secondary metabolites in fruits, vegetables, and natural products and developing appropriate databases are essential for establishing relations between diet and health.
As many as 50,000 secondary metabolites have been reported in the plant kingdom (3, 4). These compounds are not directly involved with energy production, growth, and reproduction but with defense, attraction, and protection. One source suggests (3) that there are 29,000 terpenes (derived from C5 precursors), 12,000 alkaloids (nitrogen- and sulfur-containing compounds derived from amino acids), and 8000 phenolics (derived from the shikimate pathway). These 3 main categories can be further broken down into additional subcategories. This review will be limited to phenolics, which, like the rest of the secondary metabolites, represent a tremendous analytical challenge because of their structural and chemical diversity, the complexity of the plant matrixes, and the lack of standards for the vast majority of these compounds.
The 8000 phenolic compounds are made up of phenolic acids, flavonoids, and flavonoid polymers. Phenolic acids include benzoic and cinnamic acids. Flavonoids consist of 3 basic building blocks: the flavonoid aglycones (variations of the flavonols, flavones, flavanones, flavan-3-ols, and anthocyanidins); saccharides (5 and 6 carbons); and acyl groups. The most common polymers are proanthocyanidins, which are composed of polymers of the flavan-3-ols (catechins).
HPLC coupled to diode-array detectors (DADs)5 (for UV spectra) and mass spectrometers are the most useful instruments for the nontargeted detection of phenolic compounds (5–8). More recently, ultra-HPLC (UHPLC) coupled to high-resolution multistage MS (UHPLC-DAD-HRMSn) have extended our analytical capabilities. It is now possible to putatively identify flavonoids, saccharides, acyl groups, and polymers. In some cases, MSn techniques can be used to identify isomers, but most of the time this requires further analysis by NMR spectrometry. In general, absolute identification of these compounds depends on a comparison to authentic standards. Consequently, most researchers settle for putative identification unless warranted by unusual circumstances. The Achilles heel of any of the analytical methods is the quantification of the thousands of compounds (9). As stated previously, standards are not available for most of these compounds. Even if they were available, maintaining a complete library would be a logistical and economic nightmare.
Current Status of Knowledge
Approaches to quantification include specific and nonspecific methods. Specific methods seek to separate, identify, and quantify each individual compound. The total phenolic composition is achieved by summing the concentration of the individual components (10–13). Nonspecific methods attempt to quantify the total phenolic content with a single measurement (14–20). Each approach has its strengths and weaknesses. However, to establish relations between diet and health, only specific methods are applicable.
Nonspecific methods.
The tremendous variety of phenolic compounds and lack of standards have understandably generated interest in the development of a single method that will measure the total phenolic content of a sample. Coupled with the fact that phenolic compounds are strong antioxidants, the total phenolic content has been equated to the total antioxidant capacity. A wide variety of methods have been developed for this purpose based on different chemical reactions. Examples include the Folin-Ciocalteu (interaction of a colorimetric agent with a phenolic ring) (14), oxygen radical absorbance capacity (ORAC) (15), diphenylpicrylhydrazyl (16), Trolox-equivalent antioxidant capacity (TEAC) (17), and ferric-reducing antioxidant power (FRAP) methods (18). Each of these methods has been demonstrated to measure the targeted reaction in an accurate manner and has acceptable reproducibility. However, it is not possible to relate the results of the methods to one another; i.e., no proportionality constant makes it possible to convert a TEAC number to a FRAP number (21). Unfortunately, the results for each method are affected by the compounds being measured and the sample matrix.
There is considerable controversy regarding the nutritional usefulness of measuring antioxidant capacity. Reviews by Frankel and German (22), Holst and Williamson (23), and Hollman et al. (24) have questioned the physiological relevance of the measurements as indicators of in vivo activity and their usefulness as predictors of health promotion. In addition, Finley et al. (25) observed that a delicate balance of oxidant/antioxidant reactions exists in the body and questioned the wisdom of flooding the body with antioxidants.
In 2010, industry stakeholders asked AOAC International to constitute an expert review panel (ERP) to establish method performance requirements for antioxidant capacity methods for foods. This request was based on a desire to establish a level playing field for the industry so that product antioxidant capacity numbers could be compared. The ERP reported that antioxidant capacity was an in vitro measurement that was a marketing tool (26). None of the methods measured a true antioxidant capacity. TEAC produced TEAC values, FRAP produced FRAP values, and so on. The methods were not equatable and were matrix-dependent. The ERP reported that antioxidant capacity is a complex chemical reaction and that each method measures a different aspect of the process. There is no true antioxidant capacity value.
In 2012, the USDA removed their ORAC database from the Internet. The database for 277 foods was originally released in 2007 and was expanded in 2010. According to the USDA (27), the withdrawal in 2012 was “due to mounting evidence that the values indicating antioxidant capacity have no relevance to the effects of specific bioactive compounds, including polyphenols, on human health.” The USDA further stated that “ORAC values are routinely misused by food and dietary supplement manufacturing companies to promote their products and by consumers to guide their food and dietary supplement choices.” Regarding the antioxidant capacity methods, the USDA noted that these assays are based on “discrete underlying mechanisms that use different radical or oxidant sources and therefore generate distinct values and cannot be compared directly.” The USDA did not deny the health benefits of polyphenols and acknowledged that “antioxidant molecules in food have a wide range of functions, many of which are unrelated to the ability to absorb free radicals.”
In 2013, the International Life Sciences Institute assembled a committee to draft a letter on “recommendations on reporting requirements for flavonoids in research” that was published in 2015 in the American Journal of Clinical Nutrition (28). This article covered several aspects of reporting flavonoid research. With regard to antioxidant capacity methods, the committee made the following points: 1) the methods are not specific because information on individual compounds is lost; 2) the methods are inaccurate because reducing agents are interferents; 3) the methods are not equatable because differences between methods cannot be reconciled; 4) the methods are not physiologically relevant because health benefits do not necessarily stem from antioxidant processes; and 5) the methods are not suitable for health research because nonspecific results cannot be related to health outcomes. The final point is the most important. Because the same antioxidant capacity value can be achieved in countless ways from a countless number of compounds in many different concentrations, it is impossible to relate an experimental outcome to a specific food component.
Nonspecific methods do have a place in the laboratory. They are fast and can be very useful for quality assurance under properly controlled conditions. The original application of the Folin-Ciocalteu method, for example, was for polyphenols in red wine (29). Careful research had shown the lack of interferences, and the assay provided a rapid means of judging the quality of the wine. The authors cautioned against use of the method for foods, however, explaining that any reducing agent would interfere with the accuracy of the result. Unfortunately, few researchers seem to have taken heed of this warning.
Specific methods.
As previously stated, HPLC-DAD-MS is the usual method of choice for separating and identifying phenolic compounds. Putative identification is commonplace; positive identification depends on a comparison to authentic standards. The lack of standards for positive identification is also the primary obstacle to quantification. In general, 3 different approaches have been used for quantification. The first approach requires an a priori decision as to which flavonoids are the most important and depends on those compounds for which standards are available (30, 31). This approach can successfully analyze the most common compounds—or at least those for which standards can be obtained at a reasonable price. Even those obtainable at an exorbitant price have purity issues. Waters of hydration and residual solvent are the most frequent problems despite the standards being “chromatographically” pure (32). These methods may omit the quantification of a large number of compounds and cannot be considered comprehensive.
The second approach is to hydrolyze the flavonoids to produce the aglycones (33, 34). Standards for benzoic and cinnamic acid and most common aglycones are commercially available at reasonable prices. This approach has the advantage of generating simpler chromatograms and is compatible with the assumption that only the aglycones are absorbed in the human digestive tract. The USDA supplementary database for flavonoids lists analytical data as flavonoid aglycone equivalents (35). The simplicity of the method and the availability of standards makes this approach appealing but eliminates any information regarding the original structure of the compounds.
A major obstacle to the hydrolysis method is the variable rate at which hydrolysis proceeds for the different compounds. One set of conditions is not optimal for all compounds. Merkin and Beecher (33) approached this problem by hydrolyzing samples for 5 h and removing subsamples every half hour. The 10 samples were analyzed by LC-MS, and the original aglycone concentration was computed assuming competing first-order reactions, i.e., the hydrolysis of the glycosylated flavonoid to the aglycone and the destruction of the aglycone in the acidic solution. This was a very time-consuming method; 2 hydrolysis apparatuses generated 20 samples that required >20 h on the HPLC-DAD-MS at 1 h per run. Interestingly, only the flavonols, flavones, and flavanones could be analyzed in this manner because the flavan-3-ols and anthocyanins degraded to fast under the hydrolysis conditions to be determined. Instead, the latter compounds were analyzed directly on a separate HPLC-DAD-MS run (36).
The hydrolysis method was used to analyze flavonoids in 59 fresh fruits, vegetables, and nuts over a period of 3 y (37). The relative precision of method was 25%, which was acceptable considering that the mean variability of each aglycone from 12 samplings across the country at 2 times during the year was 160%. The major disadvantages of this method are that it is extremely time- and resource-consuming and that quantitative data for specific glycosides are lost.
The third approach analyzes the phenolic compounds in their native form with use of relative response factors. Although this approach may be used with any chromatography/detector system, it is particularly effective with HPLC-DAD-MS or UHPLC-DAD-HRMSn. MS data are used for identification, but quantification is based on referencing the UV absorbance of each flavonoid peak to that of a selected internal standard (an inexpensive and commercially available compound) (10–12). Accurate quantification depends on knowing the relative response of each flavonoid compared to the internal standard. UV absorbance is particularly suitable for this purpose for 3 reasons. First, it is a very stable measurement and provides much better reproducibility than MS counts. Second, the conjugated nature of the molecular bonds of flavonoids (alternating single and double bonds) provides a theoretical basis for predicting the response factors for flavonoids for which there are no standards. Finally, a flavonoid is more likely to be nutritionally relevant if it is present at a high-enough concentration to be detected by UV absorbance as opposed to ultrasensitive MS.
In principle, the thousands of phenolic compounds will have thousands of relative response factors. However, there are some simplifying assumptions that can be used to develop UV absorbance relative response factors for families of compounds (32, 38). First, flavonoids have 2 strong UV absorbance bands based on the conjugated bond patterns of cinnamic (Figure 1) and benzoic (Figure 2) acids that are embedded in the molecules. Band 1 (320–390 nm) (Figure 3) is derived from cinnamic acid, and band 2 (250–290 nm) is derived from benzoic acid. In many cases, it is possible to add hydroxyl, methoxy, and saccharide groups to various positions on the flavonoids without disrupting the conjugated bond patterns and changing their absorbance constants. Changes in the conjugated bond patterns are easily discernable by a shift in the wavelength of maximum absorbance (λmax). Second, response factors based on molarity (rather than mass) will have molar relative response factors (MRRFs) that are constant for families of flavonoids with the same λmax.
FIGURE 1.
Hydroxycinnamic acid structure in flavonols and flavones. The trace numbers correspond to the groups of compounds listed in Tables 1 and 2. OH, hydroxyl; OMe, methoxy; Gly, glycoside. Reproduced from reference 32 with permission.
FIGURE 2.
Structures of common phenolic compounds. OH, hydroxyl: COOH, carboxyl
FIGURE 3.
Absorbance profiles (190–420 nm) for flavonols and flavones. MRRFP = best predicted molar relative response factor. Adapted with permission from reference 32.
The use of MRRF values is illustrated in Table 1 (32) for data obtained experimentally for a series of commercially available flavonol standards. In this study, rutin (quercetin-3-O-flycoside) was arbitrarily chosen as the internal standard. The MRRF column presents the response factors obtained for the samples as purchased. The MRRFD column presents the response factors for the standards after drying to remove waters of hydration and residual solvents. Unfortunately, in many cases the standards were very expensive, and insufficient mass was available for drying. The MRRFP column presents the best predicted values for routine usage. For all compounds in Table 1, the MRRF and MRRFD values agreed well. Drying produced minimal difference in the response factors, and all values were close to 1.00. For the other standards not shown (32), the MRRF and MRRFD values differed considerably, and drying produced MRRFD values closer to 1.00. This was generally the case with more highly hydroxylated and/or glycosylated compounds that were accompanied by greater waters of hydration. It can be seen that the response factors and λmax were not notably changed by the different structures and masses. Thus, any of the compounds in Table 1 could be used as an internal standard for the rest of the molecules.
TABLE 1.
MRRFs based on cinnamic acid: group 1 flavonols1
| Carbon position |
|||||||||||
| Compound | 3′ | 4′ | 5′ | 3 | 5 | 7 | MW | λmax (nm) | MRRF | MRRFD | MRRFP |
| Rutin | OH | OH | — | OGl | OH | OH | 610 | 354 | 0.98 | 1.00 | 1.00 |
| Quercetin-3-O-glucoside | OH | OH | — | OGl | OH | OH | 464 | 354 | 0.98 | 0.98 | 1.00 |
| Quercetin-3-O-galactoside | OH | OH | — | OGl | OH | OH | 464 | 354 | 1.04 | ND | 1.00 |
| Quercetin-3-O-rhamnoside (1) | OH | OH | — | OGl | OH | OH | 448 | 354 | 0.89 | ND | 1.00 |
| Quercetin-3-O-rhamnoside (2) | OH | OH | — | OGl | OH | OH | 448 | 354 | 0.79 | ND | 1.00 |
| Quercetin-3-O-arabinosylglucoside | OMe | OH | — | OGl | OH | OH | 596 | 354 | 0.99 | ND | 1.00 |
| Isorhamnetin-3-O-glucoside | OH | OH | — | OH | OH | OH | 478 | 354 | 0.95 | ND | 1.00 |
| Isorhamnetin-3-O-rutinoside | — | OH | — | OH | OH | OH | 624 | 354 | 1.00 | ND | 1.00 |
| Myricetin-3-O-rhamnoside | — | OH | OH | OGl | OH | OH | 464 | 352 | 0.83 | ND | 1.00 |
| Kaempferol-3-O-glucoside | — | OH | — | OGl | OH | OH | 448 | 348 | 0.95 | 0.96 | 1.00 |
| Kaempferol-3-O-rutinoside | — | OH | — | OGl | OH | OH | 594 | 348 | 0.88 | 0.88 | 1.00 |
| Kaempferol-3-O-robinoside-7-O-rhamnoside | — | OH | — | OGl | OH | OGl | 740 | 348 | 0.62 | ND | 1.00 |
| Syringetin-3-O-glucoside | OMe | OMe | OMe | OGl | OH | OH | 508 | 358 | 1.13 | ND | 1.00 |
| Syringetin-3-O-galactoside | OMe | OMe | OMe | OGl | OH | OH | 508 | 358 | 1.13 | ND | 1.00 |
MRRF, molar relative response factor; MRRFD, molar relative response factor based on dried standards; MRRFP, best predicted molar relative response factor; MW, molecular weight; ND, not determined; OGI, oxygen-linked glycoside; OH, hydroxyl; OMe, methoxy; λmax, wavelength of maximum absorbance. Reproduced from reference 32 with permission.
As shown in Table 2, the flavonols and flavones could be divided into 5 groups with similar absorbance profiles and λmax and MRRF values. Each group can be quantified by a single internal standard chosen from the group or a single internal standard for all groups. In the latter case, Figure 3 shows that if rutin, which is in group 4, is used as the internal standard at 354 nm, it can also be used to quantify groups 2 and 5 (MRRFP = 1.00) and groups 1 (MRRFP = 0.42) and 3 (MRRFP = 1.20). If absorbance is measured at 342 nm (where the absorbance profiles of groups 1 and 4 intersect), 354 nm (where groups 2, 4, and 5 intersect), and 370 nm (where groups 3 and 4 intersect), then an MRRFP of 1.00 can be used for all the flavonols and flavones with use of rutin as an internal standard.
TABLE 2.
MRRFs based on cinnamic acid: groups 2–51
| MRRFP | |
| Group 2 (326 nm) | |
| Chlorogenic acid | 1.00 |
| Caffeic acid | 1.00 |
| Caffeic acid methyl ester | 1.00 |
| Ferulic acid | 1.00 |
| Isoferulic acid | 1.00 |
| Sinapic acid | 1.00 |
| 1,3-Dicaffeoylquinic acid | 2.00 |
| 1,5-Dicaffeoylquinic acid | 2.00 |
| Chicoric acid (2,3-dicaffeoyltartaric acid) | 2.00 |
| Verbascoside | 1.00 |
| Isoverbascoside | 1.00 |
| Rosmarinic acid | 1.00 |
| Group 3 (336 nm) | |
| Apigenin | 1.00 |
| Apigenin 7-O-glucoside | 1.00 |
| Apigenin 7-O-rutinoside (rhoifolin) | 1.00 |
| Vitexin (8-C-glucosylapigenin) | 1.00 |
| Genkwanin (7-methoxyapigenin) | 1.00 |
| Acacetin (4′-methylapigenin) | 1.00 |
| Apigenin 7-O-neohesperidoside (isorhoifolin) | 1.00 |
| Vitexin 2"-O-rhamnoside | 1.00 |
| Isovitexin (6-C-glucosylapigenin) | 1.00 |
| Isovitexin-7-O-glu (saponarin) | 1.00 |
| Scutellarein (5,6,7,4′-tetrahydroxyflavone) | 1.00 |
| Scutellarin (scutellarein 7-O-glucuronide) | 1.00 |
| Acacetin 7-O-rutinoside (linarin) | 1.00 |
| Acacetin 7-O-neohesperidoside (fortunellin) | 1.00 |
| Hinokiflavone (3′,6-biapigenin) | 2.00 |
| Cupressuflavone (8,8′-bi-apigenin) | 2.00 |
| Group 4 (348 nm) | |
| Luteolin | 1.00 |
| Luteolin 7-O-glucoside | 1.00 |
| Diosmin (diosmetin 7-O-rutinoside) | 1.00 |
| Luteolin 7-O-glucoside | 1.00 |
| Luteolin 7,3′-O-diglucoside | 1.00 |
| Diosmetin (4′-methylluteolin) | 1.00 |
| Neodiosmin (diosmetin 7-O-neohesperoside) | 1.00 |
| Orientin | 1.00 |
| Homoorientin | 1.00 |
| Chrysoeriol (3′-methylluteolin) | 1.00 |
| 6-Methoxyluteolin | 1.00 |
| 6,7-Dimethoxy-5,3′,4′-trihydroxyflavone | 1.00 |
| Group 5 (368 nm) | |
| Quercetin | 1.00 |
| Myricetin (3,5,7,3′,4′,5′-hexahydroxyflavone) | 1.00 |
| Tamarixetin (quercetin-4′-methyl ether) | 1.00 |
| Rhamnetin (quercetin-7-methyl ether) | 1.00 |
| Isorhamnetin (quercetin-3′-methyl ether) | 1.00 |
| Isorhamnetin (quercetin-3′-methyl ether) | 1.00 |
| Syringetin (myricetin-3′,5′-dimethyl ether) | 1.00 |
| Kaempferol | 1.00 |
| Quercetin-4′-O-glucoside (spiraeoside) | 1.00 |
| 3,5,7-Trihydroxy-3′,4′,5′-trimethoxyflavone) | 1.00 |
| Kaempferol 7-O-neohesperidoside | 1.00 |
| Robinetin (3,7,3,4,5-pentahydroxyflavone) | 1.00 |
MRRF, molar relative response factor; MRRFP, best predicted molar relative response factor. Adapted from reference 32 with permission.
Similar response factors were obtained for flavan-3-ols, flavanones, chalcones, isoflavones, and stilbenes based on the conjugated bond pattern of benzoic acid (38) found in the A and B rings (Figure 2). For these compounds, 4 groups (Tables 3 and 4) were selected with common absorbance profiles. Like the flavonols and flavones, either an internal standard could be used to quantify each group or the flavonoids in all groups.
TABLE 3.
MRRFs based on benzoic acid: group 1 flavan-3-ols1
| Reference |
|||
| Compound | 40 (280 nm) | 41 (270 nm) | 32 (278 nm) |
| Catechin | 1.00 | 0.58 | 1.00* |
| Epicatechin | 1.03 | 1.00 | 0.97* |
| Gallocatechin | 0.29 | — | 0.31 |
| Epigallocatechin | 0.24 | 0.47 | 0.29* |
| Catechin gallate | 4.70 | — | 3.91 |
| Epicatechin gallate | 3.94 | — | 3.74* |
| Gallocatechin gallate | 3.56 | 2.55 | 3.41 |
| Epigallocatechin gallate | 2.66 | 2.77 | 3.14* |
| Gallic acid | — | 2.6 | 2.80* |
| Procyanidin A2 | — | — | 2.00 |
| Procyanidin B1 | — | — | 2.00 |
| Procyanidin B2 | — | — | 2.00 |
| Procyanidin C1 | — | — | 3.00* |
| Theaflavin | — | — | 5.34* |
| Theaflavin-3-O-gallate | — | — | 8.14* |
Parenthetical values for references are wavelengths at which measurements were made. *MRRF based on dried standards. MRRF, molar relative response factor. Adapted from reference 38 with permission.
TABLE 4.
MRRFs based on benzoic acid: groups 2–41
| MRRFP | |
| Group 2 (260 nm) | |
| Genistein | 1.00 |
| Genistin | 1.00 |
| Prunetin | 1.00 |
| Biochanin A | 1.00 |
| Daidzein | 0.70 |
| Daidzin | 0.70 |
| Puerarin | 0.70 |
| Formononetin | 0.70 |
| Glycetein | 0.80 |
| Glycetin | 0.80 |
| Group 3 (288 nm) | |
| Hesperitin | 1.00 |
| Hesperidin | 1.00 |
| Neohesperidin | 1.00 |
| Pinocembrin | 1.00 |
| 5-Hydroxy-7-methoxyflavanone | 1.00 |
| Naringenin | 1.00 |
| Naringin | 1.00 |
| Sukuranetin | 1.00 |
| Isosukaranetin | 1.00 |
| Didymin | 1.00 |
| Poncitrin | 1.00 |
| Eriodictyol | 1.00 |
| Eriodictyol-7-O-glucoside | 1.00 |
| Eriocitrin | 1.00 |
| Neoeriucutrin | 1.00 |
| Isoxanthohumol | 1.00 |
| Liguiritigenin | 0.60 |
| Phloretin | 1.20 |
| Phloridzin | 1.20 |
| trans-Resveratrol | 1.10 |
| trans-Resveratrol 3-glucoside | 1.00 |
| Group 4 (274 nm) | |
| Gallic acid | 1.00 |
| Hexahydrixydiphenic acid | 2.00 |
| Salicylic acid | 0.10 |
| Gentisic acid | 0.10 |
| Protocatechuic acid | 0.60 |
| Vanillic acid | 0.70 |
| 2,4,6-Trihydroxybenzoic acid | 0.30 |
| 2,3,4-Trihydroxybenzoic acid | 1.00 |
| Syringic acid | 1.30 |
Parenthetical values following group numbers are the wavelengths at which measurements were made. MRRF, molar relative response factor; MRRFP, best predicted molar relative response factor. Adapted from reference 38 with permission.
As stated earlier, use of MRRFs is particularly effective with HPLC-DAD-MS and UHPLC-DAD-HRMS. Chromatographic separation and detection with use of MS provide putative identification of all flavonoids. The correct MRRF can be selected based on Table 2 and the putative identity of the molecule or by λmax. It was observed that if the λmax of the measured compound and internal standard did not vary by >10 nm, then the accuracy of the MRRF was ±13%, and the accuracy based on dried standards, MRRFD, was ±3% (32, 38). If the difference in λmax was >10 nm, the accuracy of the MRRF was ±30% (drying made no difference). With MS identification, accuracies of ±3% can be expected.
Quantifying the proanthocyanidins, oligomers of flavan-3-ol monomers, is an interesting challenge that requires chromatographic separation and use of both HRMS and UV absorbance (39). MRRFs for flavan-3-ols measured at 278 nm are shown in Table 3. It can be seen that they are highly varied, from gallocatechin and epigallocatechin (EG) with an MRRF of 0.3 to catechin gallate and epicatechin (EC) gallate with an MRRF of 3.7. Consequently, predicting the MRRF requires knowing the composition of the oligomers. In addition, it was necessary to know whether the monomers were connected by 1 bond (B-type linkage) or 2 (A-type linkage) and whether any ECs had been replaced by EGs, epiafzelechins, epifisetinidols, or epirobinetinidols. Each produced a change in mass. The epi- form was used as a generic substitute for all the catechin/epicatechin structures.
A table of high-resolution masses (to 4 decimal places) for all the possible oligomers from 2 to 10 was computed based on the oligomer, the number of double bonds, the number of ECs replaced by epiafzelechins, the number of ECs replaced by EGs, and the number of galloyls (39). Chromatographic peaks at 278 nm were then scanned with use of selected ion-monitoring MS to determine the exact structure of the oligomers. Based on the structures, appropriate MRRFs were computed, and an internal standard was used to quantify the oligomers. With use of this approach, it was determined on a dry-weight basis that a purified grapeseed extract contained 16.6% flavan-3-ol monomers, 17.4% dimers, 14.2% trimers, and 0.5% tetramers. These weights represented the sum of the masses for 4 monomers, 27 dimers, 23 trimers, and 7 tetramers. Oligomers with a degree of polymerization >5 did not contribute substantially to the mass.
Conclusions
The thousands of secondary plant metabolites present an incredible analytical challenge. Nonspecific methods for flavonoids fail to provide information for specific compounds, suffer from numerous interferences, are not equatable to one another, and are not suitable for health research because nonspecific results cannot be related to health outcomes. However, under well-defined conditions they are useful for quality assurance. Specific methods are necessary to relate health outcomes to specific dietary components. They require relatively expensive equipment and a high level of expertise. Quantification must be based on an internal standard and experimentally verified MRRFs. This approach has been demonstrated for flavonoids and proanthocyanidins and will have to be applied to the many other secondary metabolites.
Acknowledgments
The sole author had responsibility for all parts of the manuscript.
Footnotes
Abbreviations used: DAD, diode-array detector; EC, epicatechin; EG, epigallocatechin; ERP, expert review panel; FRAP, ferric-reducing antioxidant power; HRMS, high-resolution MS; MRRF, molar relative response factor; ORAC, oxygen radical absorbance capacity; TEAC, Trolox-equivalent antioxidant capacity; UHPLC, ultra-HPLC; λmax, wavelength of maximum absorbance.
References
- 1.Steinmetz KA, Potter JD. Vegetables, fruit, and cancer prevention: a review. J Am Diet Assoc 1996;96:1027–39. [DOI] [PubMed] [Google Scholar]
- 2.Beecher GR. Overview of dietary flavonoids: nomenclature, occurrence, and intake. J Nutr 2003;133:3248S–54S. [DOI] [PubMed] [Google Scholar]
- 3.Wink, M. Introduction: biochemistry, physiology, and cological function of secondary metabolites. In: Annual plant reviews, vol 40. John Wiley & Sons, 2011.
- 4.Pichersky E, Gang DR. Genetics and biochemistry of secondary metabolites in plants: an evolutionary perspective. Trends Plant Sci 2000;5:439–45. [DOI] [PubMed] [Google Scholar]
- 5.Cuyckens F, Claeys M. Mass spectrometry in the structural analysis of flavonoids. J Mass Spectrom 2004;39:1–15. [DOI] [PubMed] [Google Scholar]
- 6.Prasain JK, Barnes S. Metabolism and bioavailability of flavonoids in chemoprevention: current analytical strategies and future prospectus. Mol Pharm 2007;4:846–64. [DOI] [PubMed] [Google Scholar]
- 7.Lin L-Z, Sun J, Chen P, Harnly J. UHPLC-PDA-ESI/HRMS/MS(n) analysis of anthocyanins, flavonol glycosides, and hydroxycinnamic acid derivatives in red mustard greens (Brassica juncea Coss variety). J Agric Food Chem 2011;59:12059–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Motilva MJ, Serra A, Macia A. Analysis of food polyphenols by ultra high-performance liquid chromatography coupled to mass spectrometry: an overview. J Chromatogr A 2013;1292:66–82. [DOI] [PubMed] [Google Scholar]
- 9.Harborne JB. The flavonoids: advances in research since 1986. J Chem Educ 1995;72:A73–A82. [Google Scholar]
- 10.Määttä-Riihinen KR, Kamal-Eldin A, Mattila PH, González-Paramás AM, Törrönen AR. Distribution and contents of phenolic compounds in eighteen Scandinavian berry species. J Agric Food Chem 2004;52:4477–86. [DOI] [PubMed] [Google Scholar]
- 11.Harbaum B, Hubbermann EM, Wolff C, Herges R, Zhu Z, Schwarz K. Identification of flavonoids and hydroxycinnamic acids in pak choi varieties (Brassica campestris L. ssp. chinensis var. communis) by HPLC-ESI-MSn and NMR and their quantification by HPLC-DAD. J Agric Food Chem 2007;55:8251–60. [DOI] [PubMed] [Google Scholar]
- 12.Obuchowicz J, Engelhardt UH, Donnelly K. Flavanol database for green and black teas utilising ISO 14502–1 and ISO 14502–2 as analytical tools. J Food Comp Anal 2011;24:411–7. [Google Scholar]
- 13.Lin L-Z, Sun J, Chen P, Monagas M, Harnly JM. UHPLC-PDA-ESI/HRMSn profiling method to identify and quantify oligomeric proanthocyanidins in plant products. J Ag food Chem 2014;62:9387–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Singleton VLORL-RRM. Analysis of total phenols and other oxidation substrates and antioxidants by means of Folin-Ciocalteu reagent. Methods Enzymol 1999;299:152–78. [Google Scholar]
- 15.Ou B, Hampsch-Woodill M, Prior RL. Development and validation of an improved oxygen radical absorbance capacity assay using fluorescein as the fluorescent probe. J Agric Food Chem 2001;49:4619–26. [DOI] [PubMed] [Google Scholar]
- 16.Polásek M, Skála P, Opletal L, Jahodár L. Rapid automated assay of anti-oxidation/radical-scavenging activity of natural substances by sequential injection technique (SIA) using spectrophotometric detection. Anal Bioanal Chem 2004;379:754–8. [DOI] [PubMed] [Google Scholar]
- 17.Böhm V, Puspitasari-Neinaber NL, Feruzzi M, Schwartz SJ. Trolox equivalent antioxidant capacity of different geometrical isomerss of α-carotene, β-carotene, lycopene, and zeaxanthin. J Agric Food Chem 2002;50:221–6. [DOI] [PubMed] [Google Scholar]
- 18.Benzie IF, Strain JJ. The ferric reducing ability of plasma (FRAP) as a measure of “antioxidant power”: the FRAP assay. Anal Biochem 1996;239:70–6. [DOI] [PubMed] [Google Scholar]
- 19.Lee J, Durst RW, Wrolstad RE. Determination of total monomeric anthocyanin pigment content of fruit juices, beverages, natural colorants, and wines by the pH differential method: collaborative study. J AOAC Int 2005;88:1269–78. [PubMed] [Google Scholar]
- 20.Prior RL, Fan E, Ji H, Howell A, Nio C, Payne MJ, Reed J. Multi-laboratory validation of a standard method for quantifying proanthocyanidins in cranberry powders. J Sci Food Agric 2010;90:1473–8. [DOI] [PubMed] [Google Scholar]
- 21.Prior RL, Xianli W, Schaich K. Standardized methods for the determination of antioxidant capacity and phenolics in foods and dietary supplements. J Agric Food Chem 2005;53:4290–302. [DOI] [PubMed] [Google Scholar]
- 22.Frankel EN, German JB. Antioxidants in foods and health: problems and fallacies in the field. J Sci Food Agric 2006;86:1999–2001. [Google Scholar]
- 23.Holst B, Williamson G. Nutrients and phytochemicals: from bioavailability to bioefficacy beyond antioxidants. Curr Opin Biotechnol 2008;19:73–82. [DOI] [PubMed] [Google Scholar]
- 24.Hollman PCH, Cassidy C, Comte B, Heinonen M, Richelle M, Richling E, Serafini M, Scalbert A, Sies H, Vidry S. The biological relevance of direct antioxidant effects of polyphenols for cardiovascular health in humans is not established. J Nutr 2011;141:989S–1009S. [DOI] [PubMed] [Google Scholar]
- 25.Finley JW, Kong A-N, Hintze KJ, Jeffery EH, Ji LL, Lei XG. Antioxidants in foods: state of the science important to the food industry. J Agric Food Chem 2011:59:6837–46. [DOI] [PubMed] [Google Scholar]
- 26.AOAC International. Stakeholder panel on strategic food analytical methods (SPSFAM): meeting minutes [Internet]. [cited 2015 Jul 29]. Available from: http://www.aoac.org/imis15_prod/aoac_docs/spsfam/mtgdocs/031313_spsfam_meeting_minutes.pdf.
- 27.USDA, Agriculture Research Service. Oxygen radical absorbance capacity (ORAC) of selected foods, release 2 [Internet]. [cited 2015 Jul 30]. Available from: www.ars.usda.gov/news/docs.htm?docid=15866.
- 28.Balentine DA, Dwyer JT, Erdman JW, Feruzzi MG, Gaine PC, Harnly JM, Kwik-Uribe CL. Recommendations on reporting requirements for flavonoids in research. Am J Clin Nutr 2015;101:1113–25. [DOI] [PubMed] [Google Scholar]
- 29.Waterhouse A. Folin-Ciocalteu micro method for total phenol in wine [Internet]. [cited 2015 Jul 29]. Available from: http://www.waterhouse.ucdavis.edu/faqs/folin-ciocalteu-micro-method-for-total-phenol-in-wine.
- 30.Brown PN, Shipley PR. Determination of anthocyanins in cranberry fruit and cranberry fruit products by high-performance liquid chromatography with ultraviolet detection: single-laboratory validation. J AOAC Int 2011;94:459–66. [PMC free article] [PubMed] [Google Scholar]
- 31.Wang Y-C, Yang Y-S. Simultaneous quantification of flavonoids and triterpenoids in licorice using HPLC. J Chromatogr B Analyt Technol Biomed Life Sci 2007;850:392–9. [DOI] [PubMed] [Google Scholar]
- 32.Lin L-Z, Harnly J, Zhang R-W, Fan X-E, Chen H-J. Quantitation of the hydroxycinnamic acid derivatives and the glycosides of flavonols and flavones by UV absorbance after identification by LC-MS. J Agric Food Chem 2012;60:544–53. [DOI] [PubMed] [Google Scholar]
- 33.Merken HM, Beecher GR. Liquid chromatographic method for the separation and quantification of prominent flavonoid aglycones. J Chromatogr A 2000;897:177–84. [DOI] [PubMed] [Google Scholar]
- 34.Olsen H, Aaby K, Borge GIA. Characterization and quantification of flavonoids and hydroxycinnamic acids in curly kale (Brassica oleracea L. Convar. acephala Var. sabellica) by HPLC-DAD-ESI-MSn. J Agric Food Chem 2009;57:2816–25. [DOI] [PubMed] [Google Scholar]
- 35.Bhagwat S, Haytowitz DB, Holden JM. USDA database for flavonoid content of selected foods, release 3. Beltsville (MD): USDA; 2011 [cited 2015 Jul 29]. Available from: http://www.ars.usda.gov/SP2UserFiles/Place/80400525/Data/Flav/Flav_R03.pdf.
- 36.Merken HM, Beecher GR. Measurement of food flavonoids by high-performance liquid chromatography: a review. J Agric Food Chem 2000;48:577–99. [DOI] [PubMed] [Google Scholar]
- 37.Harnly JM, Doherty RF, Beecher GR, Holden JM, Haytowitz DB, Bhagwat S, Gebhardt S. Flavonoid content of U.S. fruits, vegetables, and nuts. J Agric Food Chem 2006;54:9966–77. [DOI] [PubMed] [Google Scholar]
- 38.Lin L-Z, Harnly JM. Quantitation of flavanols, proanthocyanidins, isoflavones, flavanones, dihydrochalcones, stilbenes, and benzoic acid derivatives using UV absorbance after identification by LC-MS. J Agric Food Chem 2012;60:5832–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Lin L-Z, Sun J, Chen P, Zhang R-W, Fan X-E, Li L-W, Harnly JM. Profiling of glucosinolates and flavonoids in Rorippa indica (Linn.) Hiern. (Cruciferae) by UHPLC-PDA-ESI/HRMSn. J Agric Food Chem 2014;62:6618–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Wang H, Provan GJ, Helliwell K. HPLC determination of catechins in tea leaves and tea extracts using relative response factors. Food Chem 2003;81:307–12. [Google Scholar]
- 41.Pelillo M, Cuvelier ME, Biguzzi BT, Toschi G, Berset C, Lercker G. Calculation of the molar absorptivity of polyphenols by using liquid chromatography with diode array detection: the case of carnosic acid. J Chromatogr A 2004;1023:225–9. [DOI] [PubMed] [Google Scholar]