SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010 89

DIETARY SUPPLEMENTS

Determination of Total in Selected Chocolate and Confectionery Products Using DMAC

MARK J. PAYNE,WILLIAM JEFFREY HURST,andDAVI D A. STUART Hershey Center for Health and Nutrition, The Hershey Co., 1025 Reese Ave, Hershey, PA 17033 BOXIN OU and ELLEN FAN Brunswick Laboratories (USA), 50 Commerce Way, Norton, MA 02766 HONGPING JI and YAN KOU Brunswick Laboratories (China), 218 Xing Hu Rd, Suzhou Industrial Park, China 215125

A simple, specific, high-throughput colorimetric which the monomers are linked through C4÷C8 or, less method based on the reaction of frequently, C4÷C6 linkages. In the less common A-type 4-dimethylaminocinnamaldehyde (DMAC) with procyanidins, the monomers are connected through C2÷O÷C7 flavan-3-ols was developed to determine total or C2÷O÷C5 linkages. Procyanidins are widely distributed in procyanidins in selected cacao-based products. plants and are found in significant quantities in foods such as Extracts of defatted samples were dispensed into a fruits, spices, tea, wine, nuts, and cocoa (1). Of the many types of 96-well plate and reacted with DMAC. The polyphenols, flavan-3-ols, flavonols, and anthocyanidins are the absorbance of the reaction products was most abundant classes found in plants (2). The interest in measured at 640 nm and compared to polyphenol antioxidants has increased dramatically due to their commercially available B2 as a ability to scavenge free radicals and their association with a standard. The use of the 96-well plates and a plate wide range of positive health benefits, including vasodilation, reader dramatically improved sample throughput. antibacterial, anticarcinogenic, anti-inflammatory, and antiviral A standard protocol was established and used for effects (3–7). More specifically, for example, studies with cocoa further studies. The calibration was found to be and dark chocolate have shown improved cardiovascular linear from 1–100 ppm. The DMAC reagent reacted function (8, 9) and lower blood pressure (10, 11). relatively specifically to (–)-epicatechin, (+)-, epigallocatechin, gallocatechin, the The growing interest in procyanidins has generated a need for gallates of catechin, epicatechin, gallocatechin, improved quantitative analytical methods to determine and epigallocatechin, oligomeric procyanidins of procyanidins, particularly in complex food matrixes. cocoa up to n = 4, and A-type procyanidins. Little These methods are useful for several purposes, including or no reaction occurred with cyanidins and quantitation of procyanidin changes that occur during representative compounds of phenolic acids, food processing, assessment of food-to-food and flavones, flavanones, flavonols, anthocyanidins, category-to-category comparisons, product development, and isoflavones, and stilbenes. Sample precision test material characterization in clinical tests. A variety of studies were carried out on 10 different test methods have been developed to address this need, and have materials over several weeks, and yielded RSD been reviewed (12–14). Conventional methods of analysis, such values of 4.0 to 9.5%. The method was ring-tested as the widely used colorimetric assays with Folin reagent, in three laboratories using blinded test materials vanillin, or reaction with acidic butanol, can provide quantitative including cocoa beans, cocoa powder, chocolate results but have limitations. For example, the Folin-Ciocalteu liquor, dark chocolate, and milk chocolate. reagent reacts with all polyphenols, as well as some nonphenolic There was excellent agreement of the results compounds (e.g., ascorbic acid), yielding an overestimation of between laboratories. the total procyanidin content in a sample. The vanillin assay is very sensitive to reaction conditions and is not specific for procyanidins. More recently, reversed-phase (15, 16) and rocyanidins are oligomeric and polymeric compounds normal-phase LC methods (17–20) have been developed to belonging to the flavonoid class of polyphenols (Figure 1). separate individual procyanidin oligomers. While LC allows P separation and identification of individual components, These consist of flavan-3-ol monomeric units, typically catechin or epicatechin. The B-type procyanidins are polymers in standards for most procyanidin oligomers, particularly N > 2, are not commercially available, making direct quantitation of total procyanidins untenable because of fluorescence quenching of Received February 20, 2009. Accepted by AP June 9, 2009. longer-chain oligomeric and polymeric procyanidins. An added Corresponding author’s e-mail: [email protected] complication resulting from this quenching is that errors in SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 90 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010

reagent for plant tissues (30–32), and as a LC postcolumn derivatization reagent (33–38). Although some initial DMAC method validation work was carried out in plasma (28), no thorough analytical method has been published on cocoa-containing materials. This report describes method validation studies and application of the DMAC assay to the determination of procyanidins in cacao-based products.

Experimental

Figure 1. Flavanol ring structure. Reagents

(a) p-(Dimethylamino)cinnamaldehyde (DMAC, ³98% measurement of individual oligomers are compounded when purity by TLC per vendor’s certificate of analysis), large correction factors are applied and summed. (+)-catechin hydrate, (–)-epicatechin, (–)-gallocatechin, One of the more potentially useful colorimetric procedures, (–)-epigallocatechin, (–)-catechin gallate, (–)-epicatechin first applied by McMurrough and McDowell in 1978 (21) for gallate, (–)-gallocatechin gallate, (–)-epigallocatechin quantitation of flavanols in barley and hops, employs the gallate, apigenin, caffeic acid, chlorogenic acid, p-coumaric reagent 4-dimethylaminocinnamaldehyde (DMAC). The acid, daidzein, ferulic acid, gallic acid, genistein, hesperidin, ± DMAC condensation reaction with flavanols has been shown luteolin, myricetin, ( )-naringenin, quercetin hydrate, to be quite specific (12). The structural requirements for resveratrol, .—Sigma-Aldrich (St. Louis, reaction are: (1) meta-substituted dihydroxybenzene rings, such MO). as the 5,7-dihydroxy substituted A-ring of catechin and (b) Cyanidin chloride, delphinidin chloride.— epicatechin; (2) a single bond between C2 and C3; and (3)the ChromaDex (Irvine, CA). lack of a carbonyl at C4 (12). In comparison to the vanillin (c) , procyanidin B2.—Indofine Chemical assay, the DMAC assay provides improved specificity, Co. (Somerville, NJ). sensitivity, is easier to perform, and reaction products do not (d) Procyanidin B2.—Planta Analytica (Danbury, CT). degrade (12). Since the McMurrough paper, the DMAC (e) Hexanes (LC grade), reagent alcohol (histological reaction has been extended to other applications, including grade), glacial acetic acid (LC grade), hydrochloric acid, determination of flavanols in beer (22), wine (23–27), 37% (ACS Plus grade), methanol (LC grade).—Thermo plasma (28), and cranberry products (29), as a flavanol-staining Fisher Scientific (Waltham, MA).

Figure 2. Reaction of (+)-catechin, (–)-epicatechin, procyanidin B2 dimer (from three vendors), trimer, and tetramer standards with DMAC. SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010 91

Figure 3. Calibration plot comparison showing four different lots of procyanidin B2 standard from vendor Y.

Apparatus (d) 25 mm, 0.45 mm, PTFE filter disk.—Millipore (Billerica, MA) or equivalent. (a) 96-Well plate reader capable of reading at 640 nm with KC4 data acquisition and reporting software.—BioTek Operating Conditions Instruments, Inc. (Winooski, VT). The temperature of the plate reader was set to 25°Cand (b) 96-Well plate, flat-bottomed, polystyrene.—Thermo allowed to equilibrate for at least 15 min prior to use. The Fisher Scientific or equivalent. UV-Vis detector in the reader was set to 640 nm. The software (c) 12-Channel digital pipettor.—Thermo Fisher was programmed to shake the 96-well plate for 3 s, then read Scientific. the wells at 1 min intervals for 12 min.

Figure 4. Response of selected polyphenolic compounds (at 10 ppm) after reaction with the DMAC reagent. SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 92 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010

Table 1. Determination of sample precision

Test material Fat, % n Mean, mg/ga SD RSD, %

Cocoa bean (Ivory Coast) 54.4 7 14 1.11 7.8 Liquor A 53 15 16 1.27 8.0 Liquor B 54.9 8 19 0.64 3.4 Cocoa powder, natural 10.5 10 47 1.88 4.0 Alkalized cocoa powder 11 6 1.7 0.07 4.1 Dark chocolate A 35.2 10 9.4 0.34 3.7 Dark chocolate B 34.4 10 11 0.63 5.7 Dark chocolate C 42.4 6 7 0.66 9.5 Dark chocolate D 30 44 5.2 0.46 8.6 Milk chocolate 29.3 10 2.5 0.15 5.9 a Total procyanidins, B2 equivalents, on a whole product basis.

Preparation of DMAC Solution Analysis

For each full 96-well plate to be analyzed, 24 mL DMAC All blanks, standards, and samples were plated in triplicate. m solution was required. To prepare 30 mL DMAC solution, The 96-well plate was prepared by dispensing 50 L alcohol m HCl (3.0 mL) was added to 27 mL reagent alcohol, stoppered (which serves as the blank and zero-level standard), 50 L m with a glass or plastic stopper, and the solution was allowed to standard solutions (1, 10, 50, and 100 ppm), and 50 Ldiluted m cool at 4°C for 15 min. DMAC (0.0300 ± 0.001 g) was sample extracts. DMAC solution (250 L) was pipetted into weighed and transferred to the alcoholic HCl solution, and the all analysis wells, and the plate was read immediately. DMAC was dissolved completely by swirling the solution. If Calculations used for multiple runs, the DMAC solution was refrigerated between runs. The DMAC solution was prepared fresh daily. At the completion of the analyses, the software determined the maximum absorbance for each well, subtracted the Preparation of Standard Solutions absorbances of the blanks, established a standard calibration curve, and calculated mean concentrations (ppm) for each The procyanidin B2 stock solution was prepared by sample extract. Total procyanidins were calculated from these weighing 3 mg into a 5 mL volumetric flask and dissolving concentrations using the following equation: into 5 mL of 1:1 methanol–water. This stock solution was diluted using alcohol to prepare 1, 10, 50, and 100 ppm Total procyanidins solutions. The stock solution can be stored at –15°Cforat (mg/g, defatted basis, as procyanidin B2 equivalents) least 2 months. =[C´ D ´ V]/[1000 ´ S]

Sample Preparation where C is the concentration (ppm) determined from the calibration curve, D is the sample dilution, V is the sample Samples containing >5% fat require defatting prior to extraction volume (mL), and S is the sample size (g). Total analysis. Samples (about 10 g) were defatted with three procyanidins on a whole product basis can be calculated if the sequential 35 mL portions of hexanes. Samples with particles % fat is known for the sample. >2 mm or that appeared heterogeneous were ground in a coffee mill. For extraction, approximately 0.5 g sample was Results and Discussion weighed into a screw-capped test tube, and the sample mass Choice of Standard recorded. A (70 + 29.5 + 0.5, v/v/v) mixture of acetone–water–acetic acid (5.0 mL) was added to the sample, We evaluated (+)-catechin, traditionally the standard in this then the sample was mixed on a vortex mixer and sonicated at assay (12, 22, 23, 28), (–)-epicatechin, procyanidin B2 from 37 ± 3°C for 10 min. The tube was mixed on a vortex mixer three vendors (labeled X, Y, and Z), procyanidin trimer, and andallowedtostandfor25min,thenvortexedandallowedto procyanidin tetramer. Previous authors have reported varying stand for an additional 25 min. After mixing again, the sample responses with catechin, epicatechin, and polymers to was centrifuged for 5 min at 3000 rpm then filtered through a DMAC. For example, McMurrough (21) reported that 0.45 mm PTFE syringe filter. The filtered extract was diluted epicatechin afforded a higher response than catechin. We as needed with alcohol prior to analysis. The sample extracts found that, at equal concentrations, catechin and epicatechin could be stored at –15°C for 2 weeks. produced similar responses, while the three B2 standards, SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010 93

Figure 5. Comparison of total procyanidins (whole product basis) in selected cacao-based samples by the DMAC method (darker bars) and by an LC-fluorescence method (lighter bars). Error bars indicate ± 2 SDs. trimer, and tetramer yielded a decreased response (Figure 2), Stability of DMAC Solution similar to the behavior reported previously (22, 32). Because of the difference in response between monomer and Other researchers (21, 22) report that the DMAC solution oligomers, using catechin or epicatechin as the standard in this may be used for up to a week if stored in the dark. To evaluate assay would underestimate the total flavonoids in a DMAC solution stability, randomly selected defatted dark sample (13, 25), so the oligomers were viewed as better chocolate and cocoa bean sample extracts were reacted with standards. The three procyanidin B2 standards were checked DMAC daily over 9 days. The DMAC solution was kept ° for purity by LC (with fluorescence detection) and all were closed and refrigerated (4 C) between analyses, and the ° >96% pure based on peak areas, although the peak area for sample extracts were stored at –15 C between analyses. No procyanidin B2 from vendor Z was only 30% of the peak significant change in response was observed for either sample areas of vendors X and Y procyanidin B2 standards. Based on over the evaluation period (data not shown). Based on these these studies, procyanidin B2 from vendor X or Y are the results, the DMAC solution appears to be stable for at least recommended standards for this method; B2 from vendor Y 9 days if properly stored, although for the remainder of this was selected as our standard. Four different lots of B2 work the solution was prepared fresh daily. standard from vendor Y were procured over a 13-month Developing Reaction Conditions period and compared to evaluate lot-to-lot variability. Calibration curves were generated from each of the four lots A number of authors have reported that the DMAC (Figure 3), and the regression lines were compared reaction depends on two primary parameters: the solvent/acid statistically, yielding an adjusted r2 of 99.8713%. The slopes combination used to dissolve or dilute the sample and the and intercepts for the four curves were evaluated using reaction temperature (23, 33). Water has been shown to slow ANOVA (analysis of variance), which gave P = 0.1424 the reaction with DMAC, so most authors reported the (slopes) and P = 0.7460 (intercepts). Thus, the four curves can preparation of DMAC solution in methanol/HCl. Since our be considered equivalent at the 95% confidence level. laboratories conduct the assay on a 96-well plate, ethanol was used instead of methanol in an effort to reduce potential Linearity and Range evaporation of the solvent during the assay. Although the use of ethanol and isopropanol have been discouraged (24), in our To assess linearity and range, procyanidin B2 standards studies reagent alcohol performed well. were prepared at concentrations ranging from 0.03 to 400 ppm To prepare the DMAC solution, a wide range (1 to 25%) of and reacted with DMAC. The resulting absorbances HCl concentrations has been employed by other authors. were plotted versus standard concentration and a linear Lower concentrations slow the reaction rate of DMAC. The least-squares fit was applied to the data. The linear range was two laboratories that developed this procedure evaluated a established at 0.1 to 100 ppm (correlation coefficient = 0.99). range of HCl concentrations from 0.1–25%, and chose to use SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 94 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010

Figure 6. Total procyanidin results (defatted basis) of ring test conducted with three laboratories. Data shown are averages of triplicate analyses. Cocoa powder and dark chocolate B were analyzed as blind duplicates. Error bars indicate ± 2 SDs.

10% HCl to ensure the reaction mixture remains acidic and to dark chocolates. Samples were analyzed according to the afford a relatively rapid reaction rate. Using 10% HCl and method detailed above by two analysts over several weeks. ethanol as organic solvent to prepare the DMAC solution, the The mean and % RSD for each sample is displayed in Table 1. reaction is complete almost immediately, thus minimizing run times and solvent evaporation. Accuracy Specificity Only one cacao-based application of DMAC was found in the literature (36). In 2000, de Pascual Teresa et al., using The reaction of DMAC has been reported to be specific for DMAC as a postcolumn derivatizing reagent, reported that flavanols, and a few authors have evaluated the specificity of “plain chocolate” contained 7.37 mg/g total flavonoids. For DMAC with individual pure compounds (21, 22, 24, 28, 33). our initial evaluation, total procyanidins as measured by Expanding upon these earlier experiments, a series of LC (20) were determined for samples of cocoa powder (n =7 compounds from the following classes was selected to replicates), milk chocolate (n = 6), dark chocolate A (n = 18), determine the specificity of the DMAC reaction: phenolic dark chocolate B (n =10), and a sample of liquor obtained acids, flavones, flavanones, flavonols, anthocyanidins, from unfermented cocoa beans (n = 3). These samples also isoflavones, stilbenes, and flavanols. Solutions of each were analyzed (minimum of six replicates) using the DMAC compound were prepared at 1, 10, and 50 ppm, reacted with method. Results of these experiments are shown in Figure 5. DMAC, and the resulting absorbances were determined. The data demonstrate that the DMAC method generates data Results from this study for the 10 ppm concentration level are comparable to that of the LC method for this set of samples, shown in Figure 4 (results for 1 and 50 ppm solutions were except for cocoa powder. Although the DMAC result for similar and are not shown). The data show that the reaction cocoa powder is significantly higher than the LC result, the with DMAC essentially is specific to flavanols, their gallates, LC result cannot truly be considered a reference value since and certain indoles (22, 39). For compounds that were tested individual procyanidin standards are not commercially by other researchers, results agree with the DMAC specificity available. The figure also shows that SDs for the DMAC shown in this study. method are smaller than for the LC method. Precision Laboratory-to-Laboratory Repeatability Samples were selected to cover the range of processing To investigate repeatability between laboratories, a steps typically carried out in the manufacture of chocolate, collaborative study was carried out including three including several finished products. These samples included laboratories on two continents. Initially, a test sample (dark cocoa beans, chocolate liquor, cocoa powder, and milk and chocolate truffle candy) was analyzed multiple times by one SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010 95 of the laboratories to establish a mean total procyanidin value MD, http://www.nal.usda.gov/fnic/foodcomp/ using DMAC. This test sample was distributed to each of the Data/PA/PA.pdf three laboratories as a blind sample and analyzed according to (2) Robards, K., & Antolovich, M. (1997) Analyst 122, 11R–34R the method. All three laboratories reported an acceptable (3) Cortes, S.F., Valadares, Y.M., DeOliveira, A.B., Lemos, V.S., result (8, 9, and 8 mg/g versus the established value of Barbosa, M.P., & Braga, F.C. (2002) Planta Med. 68, 8 mg/g). This initial step provided some confidence that each 412–415 laboratory was able to conduct the analysis properly. Next, (4) Jayaprakasha, J.K., Selvi, T., & Sakariah, K.K. (2003) Food eight different cacao-based samples—including Jamaican Res. Int. 36, 117–122 cocoa beans, unfermented cocoa beans, natural cocoa powder, (5) Carnesecchi, S., Schnieder, Y., Lazarus, S.A., Coehlo, D., Gosse, F., & Raul, F. (2002) Cancer Lett. 175, 147–155 two milk chocolates, and three dark chocolates—were sent to (6) Garbacki, N., Tits, M., Angenot, N., & Damas, J. (2004) the participating laboratories for analysis. The cocoa powder BMC Pharmacol. 4, 1471–2210 and one of the dark chocolates were sent as blind duplicates, (7) Ursini, F., Rapuzzi, I., Toniolo, R., Tubaro, F., & giving a total of 10 samples. Each laboratory analyzed the Bontempelli, G. (2001) Method Enzymol. 335, 338–350 10 samples in triplicate, and the results of this study are shown (8) Cooper, K.A., Donovan, J.L., Waterhouse, A.L., & in Figure 6. Equivalent results were obtained on this set of Williamson, G. (2008) Br.J.Nutr.99, 1–11 samples (at the 95% confidence level), except for (9) Engler, M.B., & Engler, M.M. (2006) Nutr. Reviews 64, unfermented cocoa beans. For the cocoa powder blind 109–118 duplicate samples, the average total flavonoid result among (10) Hooper, J., Kroon, P.A., Rimm, E.B., Cohn, J.S., Harvey, I., the laboratories for the two samples was identical (37.4 mg/g). LeCornu, K.A., Ryder, J.J., Hall, W.L., & Cassidy, A. (2008) For the dark chocolate blind duplicate samples, the average Am.J.Clin.Nutr.88, 38–50 results were 4.0 and 4.3 mg/g, which were considered (11) Taubert, D., Roesen, R., Lehmann, C., Jung, N., & Schomig, equivalent at the 95% confidence level. E. (2007) J. Amer. Med. Assoc. 298, 49–60 (12) Rohr, G.E., Meier, B., & Sticher, O. (2000) Stud. Nat. Prod. Conclusions Chem. 21, 497–570 (13) Sun, B., & Spranger, M. (2005) CiLncia Téc. Vitiv. 20, 59–89 The DMAC method was shown to be rapid, easy to (14) Kelm, M.A., Hammerstone, J.F., & Schmitz, H.H. (2005) conduct, easy to standardize with a commercially available B2 Clin. Dev. Immunol. 12, 35–41 standard, and yields a single number for total procyanidins. (15) Jaworski, A.W., & Lee, C.Y. (1987) J. Agric. Food Chem. 35, 257–259 Compared to the LC method, the DMAC method is free of (16) Picinelli, A., Suzrez, B., & Mangas, J.J. (1997) Z. Lebensm. laborious standardization with multiple oligomeric standards, Unters. For. 204, 48–51 which currently are not commercially available. The (17) Rigaud, J., Escribano-Bailon, M.T., Prieur, C., Souquet, J.M., method is highly specific for polyphenols that have a & Cheynier, V. (1993) J. Chromatogr. A 654, 255–260 meta-substituted hydroxyl groups on the A-ring reacting with (18) Lazarus, S.A., Adamson, G.A., Hammerstone, J.F., & A-type and B-type procyanidin oligomers. Of the Schmitz, H.H. (1999) J. Agric. Food Chem. 47, 3693–3701 nonflavanols examined, only cyanidin and delphinidin (19) Hammerstone, J.F., Lazarus, S.A., Mitchell, A.E., Rucker, R., produced interferences, both of which were weak. The & Schmitz, H.H. (1999) J. Agric. Food Chem. 47, 490–496 method was used to quantify procyanidin levels in cocoa (20) Kelm, M.A., Johnson, J.C., Robbins, R.J., Hammerstone, beans, cocoa liquor, natural cocoa powder, and in dark and J.F., & Schmitz, H.H. (2006) J. Agric. Food Chem. 54, milk chocolate. The method demonstrated good agreement 1571–1576 with estimations of procyanidins by parallel measurements (21) McMurrough, I., & McDowell, J. (1978) Anal. Biochem. 91, using a total procyanidins LC assay. A caution with the 92–100 method is that monomeric flavanols have different extinction (22) Delcour, J.A., & de Varbeke, D.J. (1985) J. I. Brewing 91, coefficients than the oligomers. 37–40 It is recommended that this assay be used as part of a suite (23) Nagel, C.W., & Glories, Y. (1991) Am. J. Enol. Viticult. 42, of well-standardized tests including ORAC, total 364–366 polyphenols, catechin, and epicatechin monomers. (24) Vivas, N., Glories, Y., Lagune, L., Saucier, C., & Augustin, M. (1994) J. Int. Sci. Vigne Vin. 28, 319–336 Acknowledgments (25) Oszmianski, J., & Bourzeix, M. (1996) Pol. J. Food Nutr. Sci. 5, 43–50 (26) Arnous, A., Makris, D.P., & Kefalas, P. (2001) J. Agric. Food We thank the Hershey Center for Health and Nutrition for Chem. 49, 5736–5742 funding much of this work and Jessica Breeding for her help (27) Arnous, A., Makris, D.P., & Kefalas, P. (2002) J. Food in carrying out many of the experiments. Compos. Anal. 15, 655–665 (28) Kivits, G.A.A., van der Sman, F.J.P., & Tijburg, L.B.M. References (1997) Int. J. Food Sci. Nutr. 48, 387–392 (29) Fan, E., Kondo, M., Hymel, A., Zhang, L., & Ou, B. (1) U.S. Department of Agriculture (2004) USDA Database for “Quantification of A-Type Present in the Content of Selected Foods, Beltsville, Cranberries (Vacinium macrocarpon).” Presented at the SPSFAM-FLAV-11 Based from Call for Methods 12-21-2011 96 PAYNE ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 93, NO. 1, 2010

AOAC Annual Meeting, Dallas, TX, September 21–24, 2008; Poster P-W-313 (30) Feucht, W., & Treutter, D. (1990) Bull. Liaison - Groupe Polyphenols. 15, 128–131 (31) Gutmann, M. (1993) Biotech. Histochem. 68, 161–165 (32) Li, Y.-G., Tanner, G., & Larkin, P. (1996) J. Sci. Food Agric. 70, 89–101 (33) Treutter, D. (1989) J. Chromatogr. 467, 185–193 (34) Treutter, D., Santos-Buelga, C., Gutmann, M., & Kolodziej, H. (1994) J. Chromatogr. 667, 290–297 (35) de Pascual-Teresa, S., Treutter, D., Rivas-Gonzalo, J.C., & Santos-Buelga, C. (1998) J. Agric. Food Chem. 46, 4209–4213 (36) de Pascual-Teresa, S., Santos-Buelga, C., & Rivas-Gonzalo, J. (2000) J. Agric. Food Chem. 48, 5331–5337 (37) Correia, H., González-Paramis, A., Amaral, M.T., Santos-Buelga, C., & Batista, M.T. (2005) Phytochem. Anal. 16, 380–387 (38) Correia, H., González-Paramis, A., Amaral, M.T., Santos-Buelga, C., & Batista, M.T. (2006) Biomed. Chromatogr. 20, 88–94 (39) Porubsky, P.R., Scott, E.E., & Williams, T.D. (2008) Arch. Biochem. Biophys. 475, 14–17