research article anthocyanins and proanthocyanidins …...anc and pac (100 m c3g and epicatechin...

1182 Mol. Nutr. Food Res. 2013, 57, 1182–1197DOI 10.1002/mnfr.201200678

RESEARCH ARTICLE

Anthocyanins and proanthocyanidins from

blueberry–blackberry fermented beverages inhibit

markers of inflammation in macrophages and

carbohydrate-utilizing enzymes in vitro

Michelle H. Johnson1, Elvira Gonzalez de Mejia1,2, Junfeng Fan3, Mary Ann Lila4

and Gad G. Yousef4

1 Division of Nutritional Sciences, University of Illinois, Urbana, IL, USA2 Department of Food Science and Human Nutrition, University of Illinois, Urbana, IL, USA3 College of Bioscience and Biotechnology, Beijing Forestry University, Beijing, P. R. China4 Plants for Human Health Institute, NC Research Campus, North Carolina State University, Kannapolis, NC, USA

Scope: Berries are an excellent source of dietary flavonoids which have several health benefits.Methods and results: We evaluated well-characterized anthocyanins (ANCs) and proantho-cyanidins (PACs) from fermented blueberry–blackberry beverages. Wines were produced fromhighbush blueberries and blackberries grown in Illinois and blended to create ratios rang-ing from 100% blueberry to 100% blackberry. Total ANCs of the wine were strongly corre-lated to total phenolics (r = 0.99, p < 0.05) and to antioxidant capacity (r = 0.77, p < 0.05).ANC- and PAC-enriched fractions were purified from each wine blend and a phenolic profilewas generated. ANCs increased with more blackberries from 1114 to 1550 mg cyanidin-3-O-glucoside (C3G) equivalents/L. Hydrolysable tannins were identified in the PAC-enriched frac-tion. Both ANC- and PAC-enriched fractions inhibited starch-degrading enzyme �-glucosidaseand dipeptidyl peptidase-IV activity. Computational docking demonstrated that delphinidin-3-arabinoside effectively inactivated dipeptidyl peptidase-IV by binding with the lowest inter-action energy (−3228 kcal/mol). ANC and PAC (100 �M C3G and epicatechin equivalents,respectively) from blueberry–blackberry blends reduced LPS-induced inflammatory responsein mouse macrophages via the nuclear factor kappa B-mediated pathway.Conclusion: ANC- and PAC- (including hydrolysable tannins in blackberry) enriched fractionsfrom blueberry and blackberry fermented beverages are beneficial sources of antioxidants,inhibitors of carbohydrate-utilizing enzymes, and potential inhibitors of inflammation.

Keywords:

Anthocyanin / Berry / Diabetes / Hydrolysable tannins / Inflammation / Proantho-cyanidin

Received: October 14, 2012Revised: January 14, 2013

Accepted: January 23, 2013

Correspondence: Professor Elvira Gonzalez de Mejia, Depart-ment of Food Science and Human Nutrition, University of Illi-nois at Urbana-Champaign, 228 ERML, 1201 W. Gregory Avenue,Urbana, IL 61801, USAE-mail: [email protected]: +1-217-265-0925

Abbreviations: ANC, anthocyanin; C3G, cyanidin-3-O-glucoside;COX-2, cyclooxygenase-2; DPP-IV, dipeptidyl peptidase-IV;EAE, ellagic acid equivalents; HT, hydrolysable tannin; iNOS,inducible nitric oxide synthase; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium;NO, nitric oxide; ORAC, oxygen radical absorbance capacity;

1 Introduction

Blueberries and blackberries are a rich source of anthocyanins(ANCs), the water-soluble pigments which give berries theirred, blue, and purple coloring [1]. Berries also contain poly-merized phenolic compounds including hydrolysable tannins(HTs), such as ellagitannins and gallotannins, and condensedtannins known as proanthocyanidins (PACs) [2]. The con-sumption of ANCs is among the highest of all flavonoids

PAC, proanthocyanidin; PES, phenazine ethosulfate; p-p65 NF�B,phosphorylated p-65 subunit of nuclear factor kappa B; TE, troloxequivalents

C© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com

Mol. Nutr. Food Res. 2013, 57, 1182–1197 1183

due to their wide distribution in foods [3]. Positive healtheffects have been attributed to the high antioxidant capacityof these phytochemicals. ANCs have been established as hav-ing a high antioxidant capacity by scavenging free radicals andreducing oxidative stress [4]. PACs may also reduce oxidativestress and inflammation [5]. Increased phenolic intake hasbeen seen to correlate with health benefits and is linked withdecreased incidence of several diseases, including metabolicsyndrome [6] and diabetes [4].

Fermentation is known to increase the phenolic content ofberry juice products by increasing extraction from the skins,including ANCs [7], and therefore may increase the antioxi-dant capacity [8] and potential for health benefits. Moderatewine consumption as part of a healthy diet may reduce therisk of certain chronic diseases linked to inflammatory pro-cesses [6, 9]. The health benefits and sensory properties at-tributed to fermented berry products may be due to the highamounts of polyphenolic compounds present [10–12].

Type 2 diabetes occurs when the body is unable to pro-vide a proper insulin response to deal with glucose from thediet, leading to excessive blood glucose and insulin resistance.Several studies have been conducted to evaluate the reductionof absorption of glucose by berries and berry compounds toinhibit �-glucosidase and �-amylase [13–15]. One potentialdrug therapy for diabetics targets these carbohydrate-utilizingenzymes; natural sources of inhibitors of these enzymeswould be beneficial to avoid undesirable side effects that oc-cur with drug therapy [16]. Dipeptidyl peptidase-IV (DPP-IV,EC.3.4.14.5) is another current pharmaceutical target for di-abetes treatment. Inhibitors of this enzyme are emerging asantidiabetes therapy and allow for stimulation of insulin se-cretion to consequently decrease blood glucose levels [17].However, the inhibitory potential on DPP-IV by berry com-pounds needs to be further studied.

Chronic inflammation due to abnormal circulating lev-els of pro-inflammatory cytokines is associated with thedevelopment of diabetes and other metabolic diseases. Astimulated inflammatory cascade may be mediated by nu-clear factor kappa B (NF�B) signaling pathway to upreg-ulate pro-inflammatory enzymes, such as inducible nitricoxide synthase (iNOS) and cyclooxygenase-2 (COX-2). Di-etary flavonoids have been linked to anti-inflammatory path-ways [3], and many berries or berry compounds have beeninvestigated for their anti-inflammatory properties and an-tioxidant capacity or radical scavenging activity [4, 18–21].Fermented juices and wines from berries have been inves-tigated [11, 22], and in several studies they were found toinhibit inflammatory protein production or expression morethan their unfermented counterparts [12, 23–25].

The objective of this study was to evaluate well-characterized ANCs-enriched fraction and PACs-enrichedfraction present in blueberry–blackberry fermented bever-ages for their effect on markers of inflammation in RAW264.7 macrophages and carbohydrate utilization in vitro.We hypothesized that blends of ANCs and PACs presentin blueberry–blackberry fermented beverages have high an-

tioxidant capacity, the ability to reduce carbohydrate-utilizingenzymes activity, and reduce inflammation.

2 Materials and methods

2.1 Materials

Wines used were produced from highbush blueberries(Vaccinium corymbosum) and blackberries (Rubus spp.) grownat Dixon Springs Agricultural Center in Simpson, Illinois.Blueberry wine was fermented as previously described [14].Blueberry (V. corymbosum) cultivars Blue Chip, Bluecrop,Blue Haven, Blue Jay, Blueray, Bluetta, Collins, Coville, Dar-row, Earliblue, Elliot, Jersey, Late Blue, and Spartan werecollected from Dixon Springs Agricultural Center during theripening season of 2010. Blueberry juice from crushed frozenblueberries was fermented with Saccharomyces bayanus at23�C until alcohol content reached 5–6%; finished wine wasfiltered through cheesecloth. Blackberry wine was preparedat the same time using a similar procedure [14]. Blackberries(Rubus fruticosus) cultivars A-1937, A-2215, A-2241 Natchez,A-2315, APF 27, APF 40, APF 41, and Prime Jan were col-lected from Dixon Springs Agricultural Center during theripening season of 2010. Upon collection, the berries werestored frozen at −18�C until use. Wine was prepared using90.7 kg blackberries, with roughly even mixtures of all culti-vars listed above. To the blackberry lot, 180 mL SO2 solution(50 ppm), 3 g pectic enzyme solution (dissolved in water),and 10 kg sucrose solution were added. Dry wine yeast LavinEC 1118 (S. bayanus) obtained from Presque Isle Wine Cel-lars (North East, PA) was rehydrated at 40�C for 20 min withoccasional gentle stirring to remove clumps prior to inocula-tion of the juice. Per manufacturer’s recommendations, EC1118 was inoculated at 40 or 50 mg dry yeast/100 mL, andheld at 21 ± 1�C during fermentation. Blends ranging from100% blueberry to 100% blackberry were made using roomtemperature fermented wines. Blends are presented as a ra-tio of %blueberry : %blackberry. The ratios were 100:0, 70:30,50:50, 30:70, and 0:100 of blueberry:blackberry wine blends,respectively.

All solvents used for phenolic extraction were HPLC-grade and were purchased from Fisher Scientific (Pittsburg,PA). Amberlite XAD-7 was purchased from Sigma-Aldrich(St. Louis, MO), Sephadex LH-20 was purchased from GE LifeSciences (Buckinghamshire, UK). Murine macrophage RAW264.7 cell line and DMEM with L-glutamine were purchasedfrom American Type Culture Collection (ATCC) (Manassas,VA). Fetal bovine serum was from Invitrogen (Grand Island,NY). CellTiter 96 R© AQueous One Solution Proliferation assaykit (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS)/phenazine ethosul-fate (PES)) was purchased from Promega Corporation(Madison, WI). Mouse monoclonal antibodies (actin, iNOS,COX-2) and rabbit polyclonal antibody (p-p65 NF�B) wereobtained from Santa Cruz Biotechnology (Santa Cruz, CA),


1184 M. H. Johnson et al. Mol. Nutr. Food Res. 2013, 57, 1182–1197

and anti-mouse and anti-rabbit IgG horseradish peroxidaseconjugated secondary antibodies were from GE Health-care (Buckinghamshire, UK). Compounds used as stan-dards and their purity were as follows: Trolox (≥97%),glucose (≥99.5%), gallic acid (≥98%), ellagic acid (≥95%),(-)-epicatechin (≥99%), and cyanidin-3-O-glucoside (C3G)(≥99%). All chemicals and reagents were purchased fromSigma (St. Louis, MO), unless otherwise specified.

2.2 Total phenolic (TP) content

TP content was quantified using the Folin–Ciocalteu methodadapted to a microassay [26]. Briefly, to a 96-well flat bottomplate, 50 �L of 1 N Folin–Ciocalteu’s phenol reagent and 50 �Lof either sample, standard or blank were added; this mixturewas allowed to stand for 5 min before the addition of 100 �Lof 20% Na2CO3. The solution was then allowed to stand for10 min before reading at 690 nm using a Synergy 2 multiwellplate reader (Biotek, Winooski, VT). Results were expressedas milligrams ellagic acid equivalents (EAEs)/L wine.

2.3 Antioxidant capacity

Antioxidant capacity was measured by the oxygen radicalabsorbance capacity assay as described by Prior et al. [27],using 20 �L trolox standard, samples, or blank (75 mMphosphate buffer, pH 7.4), 120 �L of 116.9 nM fluorescein(final concentration 70 nm/well), 60 �L of 40 mM �,�′-azodiisobutyramidine dihydrochloride (AAPH) per well. Ablack walled 96-well plate was read at 485 and 582 nm every2 min at sensitivity 60 at 37�C using a Synergy 2 multiwellplate reader (Biotek). Results were expressed as millimolesTrolox equivalents (TE)/L.

2.4 Total ANC content

Total ANCs from wine blends were determined using theAOAC Official Method adapted from Lee et al. [28] for mi-croassay. Samples were diluted to a factor of 1:10 usingtwo buffers (pH 1.0, KCl buffer and pH 4.5, sodium acetatebuffer). Diluted solutions at each pH were transferred to a96-well plate and the absorbance was read at 520 and 700nm on a Synergy 2 multiwell plate reader (Biotek). The totalANCs were expressed as C3G equivalents in milligram perliter using the equation below:

Total monomeric anthocyanins(mg

L

)

= A × MW × D × 1000

� × PL× 0.45

where: A = (A520–A700) at pH1.0 − (A520–A700) at pH4.5; MW =449.2 g/mol for C3G; D = dilution factor determined experi-mentally with pH 1.0 buffer; PL is the constant at path length1 cm; � is 26 900 L/mol cm, the molar extinction coefficient

for C3G; and 1000 as the conversion factor from gram to mil-ligram; and 0.45 as the conversion factor from the establishedmethod [27] to the plate reader method. The AOAC OfficialMethod [28] has been standardized and validated and usesC3G to express ANC as equivalents as it is the most commonANC pigment found in nature. Similarly, (-)-epicatechin wasused to quantify PACs due to common use as a standard.The pH differential method utilizes C3G to calculate totalmonomeric ANCs and referred them as C3G equivalents,and, therefore, this was used as the quantification standardfor the HPLC peaks.

2.5 Phenolic extraction and preparation of enriched

fractions

Phenolic extraction was adapted from the method by Graceet al. [29]. To each blend, 0.1% TFA was added to stabilizethe ANCs during isolation. A rotary evaporator with tem-perature not exceeding 40�C was used to remove the alco-hol completely and most of the water from 600 mL of eachblend. The dealcoholized and concentrated beverage blendswere mixed with amberlite XAD-7 resin preconditioned withacidified water (0.1% TFA), settled at room temperature for20 min, and water was decanted. This process was repeatedthree more times until the supernatant became clear wheresugars, pectins, phenolic acids, and other impurities were re-moved from samples. XAD-7 is a moderately polar resin andis used to remove relatively polar compounds from nonaque-ous solvents, and to remove nonaromatic compounds frompolar solvents. Acidified water was used to precondition theresin and was acidified to protect the ANCs from degrada-tion. The sample with resin was loaded onto a column (30 ×3 cm) and washed thoroughly with 1.5 L acidified methanol(0.1% TFA) to elute the polyphenolic mixture from resin.The polyphenolic mixture was concentrated to remove anymethanol on a rotary evaporator (not exceeding 40�C) to yielda crude extract which contained ANC and PAC phenolicspresent in wine. This was then partitioned with ethyl acetateto remove any nonpolar compounds, and the polar eluate wasthen concentrated again on a rotary evaporator (not exceeding40�C). Samples were then stored at −80�C and subsequentlylyophilized on a Labconco freeze-dryer (Kansas City, MO).

To generate ANC- and PAC- enriched fractions, approx-imately 4 g of the freeze dried material dissolved in 10 mLof acidified water:methanol (80:20; 0.1% TFA) was loaded ona Sephadex LH-20 column (30 × 3 cm) preconditioned withwater:methanol (80:20; 0.1% TFA). With an isocratic elutionusing the same solvent ratio, fractions were collected startingwhen colored material began to elute to generate ANC frac-tions 1–4, and then 50% methanol was used to remove anyremaining ANC as fraction 5. Next, 70% acetone was usedto elute the PAC, and fractions were collected as PAC 1–4.The solvents were evaporated on a rotary evaporator (not ex-ceeding 40�C), and fractions were freeze dried, weighed, andstored at −80�C until further use. Phenolic extraction and


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1185

Figure 1. Phenolic extraction and fractionation procedure to prepare anthocyanin (ANC)- and proanthocyanidin (PAC)-enriched fractions.

fractionation procedure to prepare ANC- and PAC-enrichedfractions, indicating the methods conducted are presented inFig. 1.

ANC 1–5 and PAC 1–4 were collected from each blend.The following enriched fractions were used for further anal-ysis because they had the highest ANC or PAC concentra-tion based on HPLC analysis: ANC-3 for 100% blueberryand also for 50:50%; ANC-2, for 100% blackberry; PAC-1, for100% blueberry; PAC-4 for 50:50%; PAC-2 for 100% black-berry. The other fractions were not studied beyond HPLCanalysis.

2.6 ANC and PAC analysis

ANC analysis was conducted on a 1200 HPLC (Agilent Tech-nologies, Santa Clara, CA) using a Supelcosil LC-18 RP col-umn (250 × 4.6 mm, 5 �M) (Supelco, Bellefonte, PA). Sam-ples were prepared by dissolving 2 mg of each fraction in 1mL of methanol and filtering through 0.2 �M PTFE syringefilters (Fisher Scientific) before injection. The mobile phaseconsisted of 5% formic acid in H2O (A) and 100% methanol(B). The flow rate was held constant during the HPLC anal-ysis at 1 mL/min with a step gradient of 10%, 15%, 20%,

25%, 30%, 60%, and 10% of solvent B at 0, 5, 15, 20, 25,45, and 60 min, respectively. For PAC analyses, the same in-strumentation was used but with mobile phase consisting of95% H2O: 5% ACN with 0.1% formic acid (A) and of 95%ACN: 5% H2O with 0.1% formic acid (B). The flow rate washeld constant at 1 mL/min with a step gradient of 0%, 5%,30%, 60%, 90%, and 0% of solvent B at 0, 5, 40, 45, 50,and 55 min, respectively. ANC was detected at 520 nm andPAC at 280 nm using a DAD. Chemstation software (AgilentTechnologies) was used for both sample analysis protocol anddata processing. A reference blueberry extract (post-amberliteextract 22%) was always run as an internal standard, duringHPLC analysis, to account for any variation in the reten-tion time among runs. For qualitative peak identification ofspecific ANCs, we used pure standards that were already in-corporated into our HPLC library, and then we did compar-isons to references in the literature. MS was generated andidentification of ANCs was performed; accurate determina-tion was accomplished using our own reference library ofpeaks generated from the phenolic analysis of many blue-berry cultivars on the same equipment and under exactly thesame analytical conditions. ANC were identified based oncomparison to standards as our previously published dataindicates [20, 29].



Using the peak areas as measured by HPLC at 520 nm,total ANC were quantified from a standard curve gener-ated from 0.125, 0.25, 0.5, and 1.0 mg/mL of C3G and arepresented as C3G equivalents. In the same way, total PACwas calculated as epicatechin equivalents from the peak ar-eas measured at 280 nm. Fractions containing the highesttotal ANCs and PACs were considered ANC and PAC en-riched fractions, respectively. To further elucidate the compo-sition of the PAC, samples were analyzed with HPLC–ESI–MS using the same protocol above except at a flow rate of0.4 mL/min and Eclipse XDB-C18 RP column (250 mm ×3 mm × 5 �m, Agilent Technology, Inc., Wilmington, DE).LC/MS Accurate–Mass 6220 model (Agilent Technology,Inc.) was used for PAC analysis. Mass analyses conditionswere as follows: gas temperature at 350�C, gas flow 10 L/min,nebulizer 45 psi, ion source dual ESI, positive mode, and in-jection volume was 10 �L (2 mg/mL sample). MassHunterWorkstation Software for data acquisition version 2.B.00 andqualitative analysis version B.03.01 (Agilent Technology, Inc.)was used to run samples and analyze the total ion spectrumdata at 150–3000 m/z range.

2.7 Enzymatic studies with �-glucosidase

and �-amylase

The methods for enzymatic inhibition assays were adaptedfrom Kwon et al. [16]. For the �-glucosidase assay, to a 96-wellplate, 50 �L of sample (0.1 mg freeze dried product/mL) orpositive control (1 mM acarbose) were added to 100 �L of a1 U/mL �-glucosidase solution (in 0.1 M sodium phosphatebuffer pH 6.9) and incubated for 10 min. In total, 50 �Lof a 5 mM p-nitrophenyl-�-D-glucopyranoside solution (in0.1 M sodium phosphate buffer pH 6.9) were added brieflyto each well and incubated at 25�C for 5 min before readingthe absorbance at 405 nm. For the �-amylase assay, 500 �Lof sample or positive control (1 mM acarbose) were addedto 500 �L of 13 U/mL �-amylase solution (Type VI-B fromporcine pancreas in 0.02 M sodium phosphate buffer pH 6.9)and incubated in test tubes at 25�C for 10 min before 500 �Lof 1% soluble starch solution (dissolved in sodium phosphatebuffer and boiled for 10–15 min) was added to each tube andincubated for another 10 min. Finally, 1 mL of dinitrosalicylicacid color reagent was added and the tubes were placed in100�C water for 5 min. The mixture was diluted with 10 mLof distilled water and absorbance was read at 520 nm. Resultswere presented as percent inhibition relative to the currentdiabetes drug acarbose as the positive control, which was setat 100% inhibition.

2.8 DPP-IV inhibition and computational docking

study

Measurements of the activity and potential inhibition of DPP-IV (EC 3.4.14.5), a type II membrane glycoprotein, were done

using the DPP-IV GloTM Protease Assay (Promega, Madi-son, WI) following the manufacture’s protocol. Briefly, 50 �Lof DPP-IV GloTM reagent was added to a black-walled 96-wellplate containing 50 �L of blank, positive control, or treatment.The blank contained the vehicle control only while positivecontrol contained the vehicle and purified DPP-IV enzymeat a final concentration of 1 ng/mL. Treatments used wereenriched ANC or known inhibitor, diprotin A, and the puri-fied DPP-IV enzyme at a final concentration of 1 ng/mL. Thecontent of the wells was gently mixed using a plate shakerat 400 rpm for 30 s. DPP-IV cleavage of the provided sub-strate generated a luminescent signal by luciferase reactionwith the amount of DPP-IV enzyme available to bind Gly-Pro-aminoluciferin proportional to relative light units produced.This signal in relative light unit was measured after 30 min inan Ultra Microplate Reader (Biotek Instruments, Winooski,VT) and was compared to the inhibitory action of diprotin A.Diprotin A linear standard curve (y = −50.41x + 86.35, R2 =0.94) where y is the % maximum activity of the enzyme andx is the log10 of the concentration (�M) of known inhibitor(diprotin A) was used to calculate IC50 values.

A computational docking study to predict the affin-ity of main ANC present in the blends as C3G (Pro-tein Data Bank (PDB): 10099), malvidin-3-galactoside (PDB:671499), malvidin-3-arabinoside, delphinidin-3-glucoside(PDB: 13563), and delphinidin-3-arabinoside and diprotinA with DPP-IV enzyme (PDB: 1R9M) were conductedwith the protein-ligand docking program SwissDock(http://swissdock.vital-it.ch/, accessed July 2012). The PDBformat files of ANC and diprotin A were drawn by chem-ical drawing software Marvisketch (ver 5.10.3, ChemAxon).The OpenBabel GUI chemical expert system was used forconverting chemical file format (PDB file) of ANC (malvidin-3-galactoside and delphinidin-3-arabinoside) and diprotin Ato Mol2 files which can be read by the docking program.Docking results predicting binding modes were visualizedwith UCSF Chimera1.6.2, for displaying, animating, and an-alyzing large biomolecular systems using 3-D graphics andbuilt-in scripting. At the end, the best conformation was se-lected based on the lowest FullFitness [30]. The FullFitnessis the sum of the total energy of the system and includes asolvation term:

= E ligandintra + E recept

intra + E inter + �Gelec.solv + (� × SASA)

where E ligandintra and E recept

intra are the internal energy of ANC andDPP-IV, respectively; E inter is the interaction energy betweenthe ANC and DPPI-IV, and is equal to the sum of the van derWaals and the electrostatic interaction energies; �Gelec.solv iselectrostatic potential of solvent and calculated using the ana-lytical Generalized Born Molecular Volume (GB-MV2) modelimplemented in CHARMM; SASA is the solvent accessiblesurface area; � is 0.0072 kcal/(mol A2).


http://swissdock.vital-it.ch/

Mol. Nutr. Food Res. 2013, 57, 1182–1197 1187

2.9 Cell proliferation and treatment

Murine macrophage cell line RAW 264.7 was cultured inDMEM supplemented with 1% penicillin/streptomycin, 1%sodium pyruvate, and 10% fetal bovine serum at 37�C in 5%CO2/95% air using a CO2 Jacketed Incubator (NuAIRE DHAutoflow, Plymouth, MN). The cell proliferation assay wasconducted using the CellTiter 96 R© AQueous One SolutionProliferation assay kit containing a tetrazolium compound;MTS; and an electron coupling reagent, PES (Promega,Madison, WI). Briefly, 1 × 104 cells were seeded in a 96-wellplate and the total volume was adjusted to 200 �L with growthmedium. After 24 h incubation, the cells were treated with 5,50, or 100 �M C3G (for ANC) or epicatechin (for PAC) equiv-alents, and incubated for an additional 24 h. After this time,the growth medium was replaced with 100 �L fresh growthmedium and 20 �L MTS/PES were added to each well. Theplate was incubated for 2 h at 37�C in 5% CO2/95% air andthe absorbance was read at 515 nm in an Ultra MicroplateReader (Biotek Instruments). The percentage of viable cellswas calculated with respect to untreated cells in completemedia.

For treatments, RAW 264.7 macrophages were seeded at2 × 105 in a six-well plate, and the total volume was adjustedto 2 mL with growth medium and incubated for 48 h at37�C in 5% CO2/95% air. After 48 h incubation, the cellswere treated with 25, 50, or 100 �M C3G (for ANC) or withepicatechin (for PAC) equivalents, stimulated with 1 �g/mLLPS and incubated for an additional 24 h. After 24 h, spentmedium was collected for NO production measurement andwhole cell lysates were collected for Western blot analysis ofinflammatory markers, as follows.

2.10 NO measurement in supernatant of RAW 264.7

cell culture

Nitrite measurement was performed using Griess reaction.Briefly, 100 �L of the growth medium diluted 1:5 were platedin 96-well plate and an equal amount of Griess reagent(1% sulfanilamide and 0.1% N-1-(naphthyl)ethylenediamine-diHCl in 2.5% H3PO4) was added. The plate was incubatedfor 5 min at room temperature and the absorbance wasmeasured at 550 nm in an Ultra Microplate Reader (BiotekInstruments). The amount of NO was calculated usingsodium nitrite standard curve (y = 0.14x + 0.05, R2 = 0.997).

2.11 Western blot analysis

Treated cells were washed twice with ice-cold DMEM andtwice with ice-cold PBS and the cells were lysed with200 �L of Laemmli buffer (BioRad, Hercules, CA) con-taining 5% -mercaptoethanol. The cell lysates were son-icated for 30 s and then boiled for 5 min. Protein wasquantified using the RCDC Assay (BioRad), and 30 �g pro-

tein was loaded in 4–20% Tris–HCl gels (BioRad) for pro-tein separation. The resolved proteins were transferred to apolyvinylidene fluoride membrane (Millipore, Billerica, MA)and blocked with 3% milk in 0.1% Tris-buffered saline withTween 20 (TBST) for 1 h at room temperature. After block-ing, the membrane was washed with 0.1% TBST (five times,5 min each) and incubated with primary antibody (1:500) at4�C overnight. The membrane was washed again and incu-bated with anti-mouse or anti-rabbit IgG horseradish perox-idase conjugate secondary antibody (1:2500) for 1 h at roomtemperature. After incubation and repeated washing, the ex-pression of proteins was visualized using chemiluminescentreagent (GE Healthcare, Pittsburgh, PA) following manufac-turer’s instructions. The membrane picture was taken witha GL 4000 Pro Imaging system (Carestream Health, Inc.,Rochester, NY).

2.12 Statistical analyses

Data were expressed as means of independent duplicates withat least three replications. Statistical analysis was conductedusing the proc GLM procedures of SAS version 9.3 (SASInstitute, Inc., Cary, NC). Group mean comparisons wereconducted using LSD means and were considered to be sig-nificant at p < 0.05 based on minimum significant differencesfrom one-way analysis of variance (ANOVA) with � = 0.05.Correlations were done using Pearson’s correlation valueswith p < 0.05.

3 Results

3.1 Blackberry increased total ANCs, total phenolics,

and antioxidant capacity of wine blends

Blueberry–blackberry wine blends had total ANCs rangingfrom 76.8 (100% blueberry) to 137.2 mg C3G equivalents/L(100% blackberry); total polyphenols ranging from 2196.7(100% blueberry) to 4029.0 mg EAE/L wine (100% black-berry); and antioxidant capacity from 24.5 (70:30 blend)to 37.4 mmol TE/L (100% blackberry), as can be seen inTable 1. Both total polyphenols and total ANCs increasedconsistently with a higher percent of blackberry in the blend,with the highest values for the 100% blackberry wines. An-tioxidant capacity increased with 70% blackberry in the blend,with the highest antioxidant capacity seen again for the 100%blackberry blend.

Table 1 also shows that based on our previous study [14],and for comparative purposes, blueberry juice decreased 16.1brix, pH decreased by 2.5, and total polyphenols increasedby 625.4 mg/L EAE from week 0 to fermented wine atroom temperature. Blackberry juice decreased 17.5 brix, pHincreased only slightly (by 0.3), and total polyphenols in-creased 643.8 mg/L EAE from week 0 to fermented wineat room temperature.



Table 1. Total anthocyanins (ANCs [TA]) and total phenolics (TPs) of wine blends and antioxidant capacity (AC) of wine anthocyanins (ANC)and proanthocyanidins (PAC)1)

Blend ratio Total ANCs Total phenolics Antioxidant capacity(%blueberry (mg C3G- (mg EAE2)/L wine) (mmol Trolox Eq./L):%blackberry) equivalents/L

wine) Wine ANC3) PAC3)

100:0 76.8 ± 2.6d 2196.7 ± 42.7e 25.3 ± 2.2b 407.3 ± 48.8bc 657.0 ± 63.7a

70:30 94.3 ± 6.2cd 2611.2 ± 76.6d 24.5 ± 0.82b 667.6 ± 58.2ab 540.5 ± 74.1a

50:50 105.8 ± 6.3bc 2932.3 ± 28.5c 25.3 ± 1.0b 758.8 ± 58.2a 646.6 ± 50.0a

30:70 121.7 ± 8.7ab 3403.4 ± 50.5b 26.5 ± 0.1b 545.0 ± 63.5abc 604.7 ± 38.6a

0:100 137.2 ± 9.6a 4029.0 ± 68.4a 37.4 ± 1.4a 757.2 ± 45.9a 579.2 ± 52.3a

Correlations4)

TA_wine 1.00 0.99 0.77 0.61 −0.35TP_wine 1.00 0.84 0.59 −0.34AC_wine 1.00 0.44 −0.26AC_ANC 1.00 −0.41AC_PAC 1.00Total ANC5) 0.93 0.93 0.70 0.55 0.00Total PAC5) 0.78 0.72 0.31 0.82 0.52

1) Values are means ± SEM. Means with different letters in each column are significantly different (p < 0.05).2) EAE: ellagic acid equivalents.3) ANC and PAC at 0.1mg freeze dried product/mL.4) Correlation values are Pearson’s r-values (p < 0.05).5) Total ANC and PAC are based on HPLC data presented in Table 2.Note based on our previous results [14]: for blueberries: Wk0 (unfermented): pH, 6.3; Brix, 29.7; total polyphenols (mg/L EAE), 993.8 ± 6.9.For blackberries: Wk0 (unfermented): pH, 3.1; Brix, 29.0; total polyphenols (mg/L EAE), 2141.3 ± 101.1. Week 0 = no fermentation occurred;values are averages of independent duplicates.

3.2 Main ANCs and PACs in the blueberry

and blackberry enriched fractions

ANC identified for each blend are presented in Table 2. Chro-matographic analysis revealed up to 17 ANCs present in blue-berry and blackberry blend fermented beverages. Malvidin-3-galactoside was the main ANC present in 100% blueberry,while delphinidin-3-arabinoside was the predominant ANCin the blends containing blackberry. Total ANC, using totalpeak areas at 520 nm as quantified with HPLC in compari-son with the commercial standard C3G, increased with moreblackberry in the blend and ranged from 1112.3 (70:30 blend)to 1550.4 mg C3G equivalents/L (100% blackberry), confirm-ing the increase in total ANC seen by the wines before phe-nolic separation. The identification of ANCs and PACs werebased on their m/z, UV absorption (520 nm for ANC, and 280nm for PAC), and our previously published data on blueberryand blackberry species [19,20,29]. Total PAC were quantifiedusing peak areas at 280 nm and ranged from 560.3 (100%blueberry) to 1249.0 mg (-)-epicatechin equivalents/L (100%blackberry; Table 2). Further HPLC-ESI-MS analyses for PACrevealed complex tannin composition with highly polymer-ized compounds. In the fraction mostly composed of blue-berry matrix (100:0, 70:30, 50:50; blueberry:blackberry), in ad-dition to monomers, PAC oligomers and polymers were thepredominant compounds ranging from 2- to 14-units of cate-chin/epicatechin. This was consistent with published data oncomplex PAC in blueberry and cacao extracts [31–34]. Frac-tions derived mostly from blackberry matrix/blend (0:100,

30:70, 50:50; blueberry:blackberry) contained in addition toPAC, a complex chemical composition of HTs. Basic struc-tures for ANCs, PACs, and HTs are presented in Fig. 2.Figure 3 shows the composition of a mixture of blueberry:blackberry, with a blend of 50% each. Based on the relativeabundance of the ESI chromatogram, HTs are the predom-inant polymerized compounds in the matrix obtained frommostly blackberry wine. PACs (condensed tannins) were poly-meric forms of catechin and/or epicatechin molecules (m/z291.0234 each [M + H]+), while HTs were more complex,diverse, and with larger molecular size [20]. Based on LC–MS data across all PAC, blueberry contributed PACs, whileblackberry contributed mostly HTs (Fig. 3). The HTs in black-berry included gallic acid (m/z: 170.2129) and ellagic acid(m/z: 303.0307) which form gallotannins and ellagitannins,respectively, in diversified polymers ranging from 465.269 to2486.2511 m/z [M + H]+ present as shown in Fig. 3. Morethan 1000 HTs have been identified from simple glucogallinto pentameric ellagitannins with molecular weight rangingfrom 332 to 5000 [35]. LC–MS/MS fragmentation of HTs inblackberry samples led to the identification of several ma-jor compounds based on their m/z, UV absorption at 280nm, and published data such as ellagic acid pentoside (m/z:435.2345), ellagic acid glucose (m/z: 465.0272), galloyl–HHPglucose (m/z: 633.0259), sanguiim H-6 (m/z: 1871.2658), nob-otanin A/malabathrin B (m/z: 1736.1203), and lambertianinisomers A, C, and D (m/z: 1871.1403.5335 and 1870.9546,respectively). These identifications were in agreement withprevious publications [20,36–40]. A comprehensive structural


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1189

Table 2. Anthocyanins (ANC) identification and quantification by HPLC, and total proanthocyanidins (PACs) determined at maximumabsorption of 520 and 280 nm, respectively

RT (min) ANC ID ANCs (mg C3G equivalents/L) per blend(%Blueberry : %Blackberry)

100:0 70:30 50:50 30:70 0:100

24.44 Delphinidin-3-galactoside 10.8 11.2 9.1 6.9 nd25.98 Delphinidin-3-glucoside 26.8 51.0 34.7 12.3 8.827.33 Cyanidin-3-galactoside 6.5 6.3 nd 17.7 23.828.86 Delphinidin-3-arabinoside 11.9 nd 993.3 1079.3 1244.729.69 Cyanidin-3-O-glucoside 23.4 564.1 12.1 10.2 10.130.82 Cyanidin-3-arabinoside 140.2 16.7 8.0 nd 116.831.62 Petunidin-3-glucoside 7.3 79.6 45.7 100.9 9.632.69 Peonidin-3-glucoside 39.8 6.7 6.7 8.4 10.133.95 Petunidin-3-arabinoside 178.5 34.5 23.7 14.3 9.734.39 Malvidin-3-galactoside 292.1 32.7 11.9 38.4 6.735.19 Malvidin-3-glucoside 11.1 69.3 22.6 49.6 6.437.14 Malvidin-3-arabinoside 278.9 127.2 111.5 44.7 7.539.03 Delphinidin-6-acetyl-3-glucoside 7.3 12.6 10.3 nd nd40.55 Cyanidin-6-acetyl-3-glucoside 7.8 nd 6.1 6.1 43.941.70 Malvidin-6-acetyl-galactoside 26.7 13.6 8.0 7.7 nd42.25 Petunidin-6-acetyl-3-glycoside nd 11.9 10.9 nd nd44.40 Malvidin-6-acetyl-3-glucoside 10.8 19.4 14.6 nd ndTotal ANC 1113.7 1112.3 1403.1 1448.2 1550.4PAC total [mg

(-)-epicatechinequivalents/L]

560.3 1166.7 1244.7 1240.5 1249.0

Bold values indicate the concentration of the predominant ANC in 100% blueberry and 100% blackberry.nd, peak not detected.

elucidation of HTs in blackberry and other plant species wasrecently reviewed [41].

Freeze-dried ANC had an antioxidant capacity rangingfrom 407.3 (100% blueberry) to 758.8 mmol TE/L (50:50blend), with an increase in all blends containing blackberrycompared to 100% blueberry (Table 1), however, there was nodifference in the antioxidant capacity of PAC, having an av-erage antioxidant capacity of 605.6 mmol TE/L. Correlationsbetween components of the wines and enriched-fractions areshown in Table 1.

3.3 ANCs significantly inhibited activity

of �-glucosidase and DPP-IV

While �-amylase was not inhibited by either the ANC orthe PAC (data not shown), �-glucosidase was significantlyinhibited (Table 3). The ANC concentration needed to inhibit�-glucosidase by 50% (IC50) in C3G equivalents ranged from27.6 �M for 100% blackberry to 36.2 �M for 100% blueberry,with the 50:50 blend having an IC50 value in between the100% blends. When compared to the IC50 values for PAC,which ranged from 54.3 �M in epicatechin equivalents for100% blackberry to 115.1 �M for 100% blueberry, it can beseen that less ANC was needed to reduce �-glucosidase thanPAC. Correlations between IC50 values needed to inhibit the

enzyme are shown in Table 3. There was a strong negativecorrelation (r = −0.97, p < 0.05) between the amount of ANCrequired to inhibit �-glucosidase by 50% (IC50) and the totalANC present as determined by HPLC. Similarly, there wasa strong negative correlation (r = −0.93, p < 0.05) betweenPAC needed to inhibit �-glucosidase by 50% (IC50) and thetotal PAC present as determined by HPLC.

Compared to a standard curve of diprotin A, a known in-hibitor of DPP-IV containing an Ile-Pro-Ile sequence, ANCand PAC all had the ability to inhibit DPP-IV activity at con-centrations of 20 and 40 �M in C3G and epicatechin equiv-alents, respectively. ANC were more effective at inhibitingDPP-IV activity than the PAC, as seen by IC50 values inTable 3. Overall, the 50:50 blend was the most effective atreducing the activity of DPP-IV with an IC50 value of 4.0 �MC3G, comparable to the known inhibitor’s IC50 value of 4.3 ±1.6 mg/mL. Correlations between IC50 values needed to in-hibit the enzyme are shown in Table 3, as well as between IC50

values to inhibit both DPP-IV and �-glucosidase. A strongnegative correlation between amount of ANC (r = −0.90,p < 0.05) and IC50 values to inhibit DPP-IV indicated thatless of the compound was needed to inhibit the activity of theenzyme. Conversely, there was a strong positive correlationbetween PAC and the IC50 value to inhibit DPP-IV, indicatingthat higher PAC was required in the sample to inhibit the en-zyme activity. Results from the computational docking studyshowed that the FullFitness of the ANCs with the DPP-IV



Figure 2. Basic chemical structuresfor bioactive phytochemicals foundin blueberry and blackberry: an-thocyanins (ANCs; cyanidin, pe-onidin, delphinidin, petunidin andmalvidin), proanthocyanidin (PAC;catechin, epicatechin), and hyrolyz-able tannins (gallic acid, ellagicacid). Additions of glycoside(s) fromdifferent sugar types to anthocyani-dins results in formation of ANCslisted in Table 2. Addition of gluco-side(s) and/or polymerization of gal-lic/ellagic acids results in the forma-tion of hydrolysable tannins (HTs)shown in Fig. 3. Proanthocyaninpolymers are multiunits of catechinsand/or epicatechins.

had no obvious differences (from −3228 to −3052 kcal/mol)(Table 4). Moreover, they were as low as those of diprotin A(−3208 kcal/mol), which is a potent DPP-IV inhibitor. Theseresults suggest that ANCs contained in blueberry and black-berry have strong DPP-IV inhibiting effect. Among the fourchains in DPP-IV, chain C bound each ANC with the lowestinteraction energy. Malvidin-3-galactoside in blueberry anddelphinidin-3-arabinoside in blackberry, present in the high-est concentrations, were presumed to be the major activecomponents of blueberry and blackberry. The interactions ofchain C with the two ANCs are shown in Fig. 4. These re-sults indicate that delphinidin-3-arabinoside and malvidin-3-galactoside could effectively inhibit the activity of the enzyme.

3.4 Markers of inflammation were inhibited

in the NF�B-mediated pathway

Proliferation assays showed that none of the ANC or PACfrom 100% blueberry, 50% blueberry : 50% blackberry,

or 100% blackberry were cytotoxic to RAW 264.7 mousemacrophages at concentrations up to 100 �M in C3G orepicatechin equivalents, respectively. None of the ANC orPAC (100 �M C3G or epicatechin equivalents) increasedproduction of any of the inflammatory proteins tested whennot stimulated with LPS. All three (100:0, 50:50, and 0:100)blueberry–blackberry blend ANC at 100 �M were able to lowertotal p-p65 NF�B subunit following LPS-stimulation by 68.9,88.6, and 80.8%, respectively (Fig. 5A). ANC from the blendscontaining 50% or 100% blackberry were able to significantlyinhibit the activation of the p-p65 NF�B subunit at 50 �M by88.6 and 72.0%, respectively. PAC from the 100% blueberryand the 50:50 blends were also able to significantly decreasethe p-p65 NF�B subunit at 100 �M epicatechin equivalentsby 55.7 and 40.0%, respectively; however, PAC from the 100%blackberry blend did not significantly decrease this inflamma-tory protein (Fig. 5A). This can be attributed to the fact that,unlike blueberry with more polymeric PACs or condensedtannins, blackberry contains HTs. COX-2 expression was de-creased by all ANC at 100 �M C3G equivalents; 52.4% for


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1191

Figure 3. HPLC–ESI–MS chromatograph for PAC from 50:50 blend of blueberry:blackberry. The large bold numbers refer to average degreeof polymerization for proanthocyanidins (PACs), consisting of dimers (2 units of catechin/epicatechin) to 14-mers, with peaks contained twobold numbers indicated the presence of a doubly charged ion and the presence of both degrees of polymerization (i.e., 7 of 14 heptamerand 14-mers). Mass ions tagged with HT refer to polymerized hydrolysable tannins (HTs) contributed only by blackberry. HTs included gallicacid (m/z: 170.2129) and ellagic acid (m/z: 303.007) that formed gallotannins and ellagitannins in diversified polymers shown in LC–MSspectrum ranging from of 465.268 to 2486.2511 m/z [M + H]+.

100:0, 67.8% for 50:50, and 72.2% for 0:100 (Fig. 5B). PACfrom 100% blackberry and the 50:50 blends were also able todecrease COX-2 expression significantly by 64.3 and 65.2%,respectively. PAC at 100 �M epicatechin equivalents from the100% blueberry blend decreased COX-2 expression by 36.8%,however, this decrease was not significant.

iNOS expression was decreased by ANC with the 100%blueberry and 50:50 blends at 100 �M C3G equivalents,with ANC from the 100% blueberry the most effective with75.7% inhibition compared to 40.2% for 50:50 (Fig. 6A). PACfrom all blends were able to decrease iNOS expression in

mouse macrophages following LPS-stimulation at 100 �Mepicatechin equivalents, with the 50:50 blend the most ef-fective to decrease iNOS expression by 57.4% compared to53.0% inhibition by 100:0 and 42.3% by 0:100. iNOS proteinproduces the inflammatory marker NO when active, so NOproduction was also measured. As seen in Fig. 6B, ANC at100 �M were able to decrease NO production with black-berry in the blend by 37.9% for 50:50 and 48.1% for0:100; however, the 100% blueberry ANC was unable toshow a significant decrease in NO production. PAC from100:0, 50:50, and 0:100 at 100 �M epicatechin equivalents

Table 3. Determination of IC50 values (mean ± standard error) of anthocyanins (ANC) and proanthocyanidins (PAC) needed to inhibitenzymes involved in diabetes management.1)

Enriched 100% 50%:50% 100% Correlations2)

fraction Blueberry Blackberry �-glu_ �-glu_ DPP-IV_ DPP-IV_ANC PAC ANC PAC

�-Glucosidase IC50 ANC- 36.2 ± 13.2b 33.7 ± 9.8b 27.6 ± 9.7b 1.00 0.90 0.98 −0.86values (�M) PAC- 115.1 ± 28.5a 71.3 ± 28.9ab 54.3 ± 30.8ab 1.00 −0.99 −1.00

Dipeptidyl ANC- 15.9 ± 0.7a 4.2 ± 1.5b 5.5 ± 0.4b 1.00 −0.95peptidase-IV PAC- 16.6 ± 1.1a 19.3 ± 1.1a 18.2 ± 3.2a 1.00(DPP-IV) IC50values (�M)

1) IC50 values were determined from at least two independent duplicates done in triplicate. Values are means ± SEM. Means with differentletters in each row are significantly different for each enzyme (p < 0.05). Positive control of inhibition for �-glucosidase was acarbosehaving an IC50 value of 500 �M. The positive control of inhibition for DPP-IV was diprotin A (Ile-Pro-Ile) at 0.1 mg/mL, and had an IC50 valueof 4.3 ± 1.6 mg/mL.2) Correlation values are Pearson’s r-values (p < 0.05).



Table 4. Interactions of chains of dipeptidyl peptidase-IV (DPP-IV) and anthocyanins (ANCs)

Chain of FullFitness SimpleFitnessa) H bond Distance of �G (kcal)DPP-IV (kcal/mol) (kcal/mol) position H bond (A)

Cyanidin-3-O-glucoside A −3096.22 83.08 − − −7.34B −3076.18 84.01 PHE95 HN−C3G O8 2.135 −7.68C −3118.97 85.40 TRP62 NH−C3G O1 2.563 −7.26D −3051.96 78.58 TRP62 O−C3G H7 2.032 −7.83

Malvidin-3-galactoside A −3079.03 124.70 − − −7.23B −3063.75 115.45 ASN74 O−M3GA H11 2.622 −7.79C −3106.89 115.08 PHE461 HN−M3GA O10 2.275 −7.87

ILE107 O−M3GA H 2.691D −3033.99 116.42 TRP305 NH−M3GA O4 2.145 −7.59

Malvidin-3-arabinoside A −3094.00 116.12 GLY99 HN−M3A O7 2.151 −7.63B −3075.47 110.77 − − −7.88C −3118.97 113.62 ARG669 O−M3A H4 2.836 −7.73D −3046.40 113.68 − − −7.45

Delphinidin-3-glucoside A −3102.27 78.94 − − −7.59B −3081.15 80.25 − − −7.47C −3125.95 82.11 HIS533 HN−D3G O9 2.168 −7.64

ASN506 O−D3G H 2.114LYS41 HN−D3G O 2.385

D −3052.84 79.68 − − −7.48Delphinidin-3-arabinoside A −3202.45 35.04 ASN74 O−D3A H8 2.043 −7.64

ILE102 HN−D3A O4 2.467ASN92 O−D3A H14 2.150

B −3183.53 30.69 ASN74 O−D3A H8 1.763 −8.38ILE102 HN−D3A O9 2.194

C −3227.73 36.22 ILE107 HN−D3A O10 2.223 −7.62TRP62 O−D3A H14 1.990ILE63 O−D3A H9 2.338

D −3155.24 39.17 − − −7.03Diprotin A A −3183.60 −7.07 LYS71 HZ3−DA O4 2.618 −7.62

B −3161.06 −8.22 LEU90 O−DA H12 2.582 −8.25PHE95 O−DA H22 2.643HIS100 O−DA H23 2.428ILE102 HN−DA O4 1.874

C −3208.29 −9.23 LYS122 O−DA H22 2.678 −7.72GLU191 OE2−DA H12 1.875

D −3135.71 −11.14 − − −7.17

C3G, M3GA, M3A, D3G, D3A, and DA represent cyanidin-3-O-glucoside, malvidin-3-galactoside, malvidin-3-arabinoside, delphinidin-3-glucoside, delphinidin-3-arabinoside, and diprotin A, respectively. Other abbreviations indicate the three-letter amino acid code in thebinding site.a) The SimpleFitness is equal to the total energy of the system calculated with the CHARMM22 molecular mechanics force field, andneglects the effect of solvent known to have an important contribution to the binding free energy.

also decreased NO production by 16.4, 4.8, and 14.7%, re-spectively; however, these decreases were not significant(p > 0.05).

4 Discussion

Fermented berry beverages have been established as hav-ing an increased phenolic content and higher antioxidantcapacity than their nonfermented counterparts [7–9,23]. Ben-eficial effects of berries in the diet have been attributedto their high phenolic content, especially their ANC con-tent [40]. Wines were tested for their total polyphenols, to-tal ANC content, and antioxidant capacity, and these lev-

els increased with more blackberry in the blend associ-ated with higher ANC concentration. To our knowledge,this is the first reported analysis of ANCs from fermentedblueberry–blackberry blends; however, the results fromblueberry and blackberry wines were comparable to previ-ously reported values [24]. In this study, we show that theANC profile from fermented berry beverages may be alteredfrom profiles previously generated for blackberry fruits [20] oreven phenolic-rich blueberry extracts [29]. Malvidin containstwo methoxy side chains on its B ring, while delphinidinhas two hydroxyl groups, which may contribute to the higherantioxidant capacity seen for the blackberry blends. Delphini-din (45.0–74.9 mg/100 g fresh weight, FW) and malvidin(37.1–62.2 mg/100 g FW) were the predominant ANCs in the


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1193

Figure 4. Computational visualization of the most optimal docking conformation of (A) delphinidin-3-arabinoside, (B) malvidin-3-galactoside, and (C) diprotin A, with chain C of the DPP-IV enzyme.



Figure 5. Effect of anthocyanin(ANC; cyanidin-3-O-glucosideequivalents) and proanthocyanidin(PAC; epicatechin equivalents) at50 and 100 �M on NF�B (A) andCOX-2 expression (B). Westernblots are shown above with p-p65subunit of NF�B detected at 65 kDa,COX-2 at 70–72 kDa, and actin at43 kDa from whole cell lysates.Values are relative expressionof protein of interest over actin;means ± SEM with different lettersare significantly different (n = 4,p < 0.05).

varieties tested in highbush and lowbush blueberries [42],similar to the findings from the fermented products.

During wine processing from berries, phenolic com-pounds may be effectively extracted into the wines;delphinidin-3-glucoside has been identified as a major com-ponent in red wine and had the most variation among allANCs [43]. Although C3G has been identified as the majorANC in different blackberry species, cyanidin 3-rutinosidewas the major component in the wines from Andean black-berry (Rubus glaucus Benth.) [44]. These results contrast withthe major ANC identified in our study as delphinidin-3-arabinoside, though many unknown compounds in the pre-vious study were only tentatively identified. Authors agreethat fermentation causes an absolute and relative change inANC content, allowing for new pigments to be generated

or formed during the fermentation process or during winematuration. While the process of ANC transformation is notclear, it is supported by the changes seen in wines duringaging as native blackberry ANCs are transformed during andafter alcoholic fermentation.

Our results showed inhibition of �-glucosidase activ-ity; other studies on red and black currants and red andgreen gooseberries have shown strong inhibition of intestinalglucose absorption by inhibiting �-glucosidase activity [13].ANCs present in blueberry fruits and their fermented prod-ucts appear to be more potent at inhibiting �-glucosidase thanthe PACs. While a previous study in our laboratory evaluatedblueberry wines and found strong inhibition of �-amylasein vitro [14], we believe that during the phenolic extractionprocess, phenolic acids were removed indicating these may


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1195

Figure 6. Effect of anthocyanin(ANC; cyanidin-3-O-glucoside equi-valents) and proanthocyanidin(PAC; epicatechin equivalents)at 50 and 100 �M on (A) iNOSexpression. Western blots areshown above with iNOS detectedat 130 kDa and actin at 43 kDafrom whole cell lysates. Values arerelative expression of protein ofinterest over actin. (B) NO assaypresents data from treatments at100 �M cyanidin-3-O-glucoside(C3G) equivalents (ANC) or epicat-echin equivalents (PAC) and wasmeasured from spent medium.Means ± SEM with different lettersare significantly different (n = 4,p < 0.05).

be the bioactive compounds with ability to inhibit �-amylaseactivity rather than the ANC or PAC. Grussu et al. [45] demon-strated that it is not the ANCs that are required to inhibit�-amylase because both red and yellow raspberries showedsimilar inhibitory capacity.

A study of rowanberry PAC-rich fraction had a lowlevel of inhibition compared to the whole extract [46], sug-gesting that tannin and ellagitannin components are notinfluential or are poor inhibitors of �-glucosidase. Our re-sults indicate that there was some inhibition of the enzymeby the PAC-enriched fraction, but the ANC-enriched frac-tion was more potent. The efficacy of the ANCs to inhibitDPP-IV enzyme activity at a rate comparable to a known in-hibitor (diprotin A) indicated that ANCs may be able to actas naturally occurring DPP-IV inhibitors of DPP-IV. If DPP-IV is inactivated by drugs or natural compounds, the degra-dation of glucagon-like peptide-1 and glucose-dependent in-sulinotropic polypeptide is prevented, and these hormonesare able to inhibit glucagon release and stimulate insulinrelease, respectively, thereby lowering blood glucose levels[17]. All of the ANCs tested showed similar Fullfitness ener-gies from the computational modeling study; therefore, theyhave no major difference in anti-DPP-IV activity. Due to thehigher amounts of malvidin-3-galactoside in blueberry anddelphinidin-3-arabinoside in blackberry, these ANC are likely

the major active components. To confirm these results, theSwissDock computational docking study demonstrated thatdelphinidin-3-arabinoside and malvidin-3-galactoside wereable to interact with DPP-IV through ligand interaction, andtherefore potentially inactivate the activity of the enzyme.The most optimal conformation happened in chain C ofDPP-IV indicated by the lowest interaction energies of −3228kcal/mol (for delphinidin-3-arabinoside) and −3107 kcal/mol(for malvidin-3-galactoside), which are comparable to the in-teraction energy of the known inhibitor diprotin A with theenzyme of −3208 kcal/mol. Due to the lowest interaction en-ergy of chain C, we concluded that the chain C was the majorbinding site of these ANCs.

Our results agree with previous studies that reported theability of berries to reduce inflammation through decreasedexpression of enzymes iNOS and COX-2 [19, 20]. The up-stream mediator in the NF�B-mediated pathway, the p-p65subunit of NF�B that lead to the induction of these inflamma-tory markers was inhibited by ANC 50:50% blend and 100%blackberry. The strong anti-inflammatory effect was consid-ered to be exerted by a combination of various polyphenolcompounds rather than any single compound.

Berry extracts can inhibit �-glucosidase at concentrationsless than 50 �g GAE/mL, an amount that can be eas-ily reached in the gastrointestinal tract after consumption



of berries or juices [47]. Additionally, a previous study re-ports that a moderate drink of 140 mL of wine made fromhighbush blueberry would correspond to an intake of 2.42mmol of TE of Oxygen Radical Absorbance Capacity beta-Phycoerythrin/day [43].

In conclusion, ANC and PAC at 100 �M C3G/epicatechinequivalents obtained from blueberry–blackberry blends af-fected LPS-induced inflammatory response in RAW 264.7mouse macrophages. These ANC and PAC reduced the phos-phorylation of the p65-subunit NF�B, decreased iNOS andCOX-2 expression, and decreased NO production. Along withthe high antioxidant capacity and phenolic content and abil-ity to decrease enzyme activity related to improved glucoseutilization, bioactive ANC and PAC compounds present infermented berry beverages may act synergistically to reducerisk factors for chronic inflammatory diseases, especially type2 diabetes.

Thanks to Dixon Springs Agricultural Center in Simpson,IL, for providing materials for the fermentation. Special thanksto Yan Wei and Christina Alcaraz for their initial work withcomputational docking, and to the State scholarship fund for J.F.from Beijing Forestry University and China Scholarship Council,China. The Office of Research provided funds for this research.

The authors have declared no conflict of interest.

5 References

[1] Miguel, M. G., Anthocyanins: antioxidant and/or anti-inflammatory activities. J. App. Pharm. Sci. 2011, 01, 07–15.

[2] Cuevas-Rodriguez, E. O., Dia, V. P., Yousef, G. G., Garcia-Saucedo, P. A. et al., Inhibition of pro-inflammatory re-sponses and antioxidant capacity of Mexican blackberry(Rubus spp.) extracts. J. Agric. Food Chem. 2010, 58, 9542–9548.

[3] Wu, X., Schauss, A. G., Mitigation of inflammation withfoods. J. Agric. Food Chem. 2012, 60, 6703–6717.

[4] Tsuda, T., Dietary anthocyanin-rich plants: biochemical basisand recent progress in health benefits studies. Mol. Nutr.Food Res. 2012, 56, 159–170.

[5] Lee, Y. A., Kim, Y. J., Cho, E. J., Yokozawa, T., Ameliorativeeffects of proanthocyanidin on oxidative stress and inflam-mation in streptozotocin-induced diabetic rats. J. Agric. FoodChem. 2007, 55, 9395–9400.

[6] Burton-Freeman, B., Postprandial metabolic events and fruit-derived phenolics: a review of the science. British J. Nutr.2010, 104, S1-S14.

[7] Su, M. S., Chien, P. J., Antioxidant activity, anthocyanins,and phenolics of rabbiteye blueberry (Vaccinium ashei) fluidproducts as affected by fermentation. Food Chem. 2007, 104,182–187.

[8] Martin, J. L., Matar, C., Increase of antioxidant capacity ofthe lowbush blueberry (Vaccinium angustifolium) during fer-mentation by a novel bacterium from the fruit microflora. J.Sci. Food Agric. 2005, 85, 1477–1484.

[9] Walzen, R. L., Review: wine and health—state of proofsand research needs. Inflammopharmacology 2008, 16,265–271.

[10] Vuong, T., Martin, L., Matar, C., Antioxidant activity of fer-mented berry juices and their effects on nitric oxide and tu-mor necrosis factor-alpha production in macrophages 264.7gamma NO(-) cell line. J. Food Biochem. 2006, 30, 294–268.

[11] Juan, C., Jianquan, K., Junni, T., Zijian, C. et al., The profile inpolyphenols and volatile compounds in alcoholic beveragesfrom different cultivars of mulberry. Food Chem. 2012, 77,C430-C436.

[12] Lim, J. W., Hwang, H. J., Shin, C. S., Polyphenol compoundsand anti-inflammatory activities of Korean black raspberry(Rubus coreanus Miquel) wines produced from juice sup-plemented with pulp and seed. J. Agric. Food Chem. 2012,60, 5121–5127.

[13] Da Silva Pinto, M., Kwon, Y. I., Apostolidis, E., Lajolo, M. F.et al., Evaluation of red currants (Ribes rubrum L.), black cur-rants (Ribes nigrum L.), red and green gooseberries (Ribesuva-crispa) for potential management of type 2 diabetes andhypertension using in vitro models. J. Food Biochem. 2010,34, 639–660.

[14] Johnson, M. H., Lucius, A., Meyer, T., de Mejia, E. G., Cultivarevaluation and effect of fermentation on antioxidant capac-ity and in vitro inhibition of �-amylase and �-glucosidaseby highbush blueberry (Vaccinium corombosum). J. Agric.Food Chem. 2011, 59, 8923–8930.

[15] Rojo, L. E., Ribnicky, D., Logendra, S., Poulev, A. et al.,In vitro and in vivo anti-diabetic effects of anthocyanins fromMaqui berry (Aristotelia chilensis). Food Chem. 2012, 131,387–396.

[16] Kwon, Y. I., Apostolidis, E., Shetty, K., Inhibitory potentialof wine and tea against �-amylase and �-glucosidase formanagement of hyperglycemia linked to type 2 diabetes.J. Food Biochem. 2008, 32, 15–31.

[17] Parmar, H. S., Jain, P., Chauhan, D. S., Bhinchar, M. K. et al.,DPP-IV inhibitory potential of naringin: an in silico, in vitroand in vivo study. Diab. Res. Clin. Pract. 2012, 97, 105–111.

[18] DeFuria, J., Bennett, G., Strissel, K. J., Perfield, J. W., II et al.,Dietary blueberry attenuates whole-body insulin resistancein high fat-fed mice by reducing adipocyte death and itsinflammatory sequelae. J. Nutr. 2009, 139, 1510–1516.

[19] Schreckinger, M. E., Wang, J., Yousef, G., Lila, M. A. et al.,Antioxidant capacity and in vitro inhibition of adipogenesisand inflammation by phenolic extracts of Vaccinium flori-bundum and Aristotelia chilensis. J. Agric. Food Chem. 2010,58, 8966–8976.

[20] Cuevas-Rodriguez, E. O., Yousef, G. G., Garcia-Saucedo,P. A., Lopez-Medina, J. et al., Characterization of antho-cyanins and proanthocyanidins in wild and domesticatedMexican blackberries (Rubus spp.). J. Agric. Food Chem.2010, 58, 7458–7464.

[21] Xie, C., Kang, J., Ferguson, M. E., Nagarajan, S. et al., Blue-berries reduce pro-inflammatory cytokine TNF-� and IL-6production in mouse macrophages by inhibiting NF-�B ac-tivation and the MAPK pathway. Mol. Nutr. Food Res. 2011,55, 1587–1591.


Mol. Nutr. Food Res. 2013, 57, 1182–1197 1197

[22] Riso, P., Klimis-Zacas, D., Del Bo’, C., Martini, D.et al., Effect of a wild blueberry (Vaccinium angus-tifolium) drink intervention on markers of oxidativestress, inflammation and endothelial function in hu-mans with cardiovascular risk factors. Eur. J. Nutr.2012. doi: 10.1007/s00394-012-0402-9.

[23] Wang, J. Z., Yousef, G. G., Rogers, R. B., Gonzalez de Mejia,E. et al., in: Bhimanagouda S. Patil, Guddadarangavvana-hally K. Jayaprakasha, Kotamballi N. Chidambara Murthy,Navindra P. Seeram (Eds.), ACS Books Emerging Trends inDietary Components for Preventing and Combating Disease,Chapter 6, pp. 95–116.

[24] Jeong, J. H., Jung, H., Lee, S. R., Lee, H. J. et al., Anti-oxidant, anti-proliferative, and anti-inflammatory activitiesof the extracts from black raspberry fruits and wine. FoodChem. 2010, 123, 338–344.

[25] Vuong, T., Martineau, L. C., Ramassamy, C., Matar, C.et al., Fermented Canadian lowbush blueberry juice stim-ulates glucose uptake and AMP-activated protein kinase ininsulin-sensitive cultured muscle cells and adipocytes. Can.J. Physiol. Pharmacol. 2007, 85, 956–965.

[26] Heck, C. I., Schmalko, M., Gonzalez de Mejia, E., Effect ofgrowing and drying conditions on the phenolic compositionof mate tea (Ilex paraguariensis). J. Agric. Food Chem. 2008,56, 8394–8403.

[27] Prior, R., Hoang, H., Gu, L., Wu, X. et al., Assays for hy-drophilic and lipophilic antioxidant capacity (oxygen radicalabsorbance capacity (ORAC(FL)) of plasma and other bio-logical and food samples. J. Agric. Food Chem. 2003, 51,3273–3279.

[28] Lee, J., Durst, R. W., Wrolstad, R. E., Determination of totalmonomeric anthocyanin Pigment content of fruit juices, bev-erages, natural colorants, and wines by the pH differentialmethod: collaborative study. J. AOAC Int. 2005, 88, 1269–1278.

[29] Grace, M., Ribnicky, D., Kuhn, P., Poulev, A. et al., Hypo-glycemic activity of a novel anthocyanin-rich formulationfrom lowbush blueberry, Vaccinium angustifolium Aiton.Phytomedicine 2009, 16, 406–415.

[30] Grosdidier, A., Zoete, V., Michielin, O., EADock: docking ofsmall molecules into protein active sites with a multi objec-tive evolutionary optimization. Proteins 2007, 67, 1010–1025.

[31] Alwerdt, J., Seigler, D. S., Gonzalez de Mejia, E., Yousef, G.G. et al., The influence of alternative liquid chromatographytechniques on chemical complexity and bioactivity of sepa-rated proanthocyanidin mixtures. J. Agric. Food Chem. 2008,56, 1896–1906.

[32] Gu, L., Kelm, M., Hammer stone, J. F., Beecher, G.et al., Fractionation of polymeric procyanidins from low-bush blueberry and quantification of procyanidins in se-lected foods with and optimized normal-phase HPLC-MS flu-orescent detection method. J. Agric. Food Chem. 2002, 50,4852–4860.

[33] Kelm, M. A., Johnson, J. C., Robbins, R. J., Hammerstone, J.F. et al., High-performance liquid chromatography separationand purification of cacao (Theobroma cacao L.) procyanidinsaccording to degree of polymerization using a diol stationaryphase. J. Agric. Food Chem. 2006, 54, 1571–1576.

[34] Guyot, S., Marnet, N., Drilleau, J., Thiolysis-HPLC charac-terization of apple procyanidins covering a large range ofpolymerization states. J. Agric. Food Chem. 2001, 49, 14–20.

[35] Serrano, J., Puupponenpimi, R., Dauer, A., Aura, A. M.et al., Tannins: current knowledge of food sources, intake,bioavailability and biological effects. Mol. Nutr. Food Res.2009, 53, 310–329.

[36] Hager, T. J., Howard, L. R., Liyanage, R., Lay, J. O. et al., Ellag-itannin composition of blackberry as determined by HPLC-ESI-MS and MALDI-TOF-MS. J. Agric. Food Chem. 2008, 56,661–669.

[37] Mullen, W., Yokota, T., Lean, M. J., Crozier, A., Analysis ofellagitannins and conjugates of ellagic acid and quercetinin raspberry fruits by LC-MSn. Phytochemistry 2003, 64,617–624.

[38] Yoshida, T., Nakata, F., Hosotani, K., Nitta, A. et al., Dimerichydrolysable tannins from Melastoma malabathricum. Phy-tochemistry 1992, 31, 2829–2833.

[39] Hassimoto, A. N. M., Mota, R. V., Cordenunsi, B. R.,Lajolo, F. M., Physico-chemical characterization andbioactive compounds of blackberry fruits (Rubus sp.)grown in Brazil. Cienc. Technol. Aliment. 2008, 28,702–708.

[40] Mertz, C., Cheynier, V., Guanata, Z., Bart, P., Analysis of phe-nolic compounds in two blackberry species (Rubus glaucusand Rubus adenotrichus) by high-performance liquid chro-matography with diode array detection and electrosprayion trap mass spectrometry. J. Agric. Food Chem. 2007, 55,8616–8624.

[41] Arapitsas, P., Hydrolyzable tannin analysis in food. FoodChem. 2012, 135, 1708–1717.

[42] Rodriguez-Mateos, A., Cifuentes-Gomez, T., Tabatabaee, S.,Lecras, C. et al., Procyanidin, anthocyanin, and chlorogenicacid contents of highbush and lowbush blueberries. J. Agric.Food Chem. 2012, 60, 5772–5778.

[43] Sanchez-Moreno, C., Guohua, C., Ou, B., Prior, R. L.,Anthocyanin and proanthocyanidin content in selectedwhite and red wines. Oxygen radical absorbance ca-pacity comparison with nontraditional wines obtainedfrom highbush blueberry. J. Agric. Food Chem. 2003, 51,4889–4896.

[44] Arozarena, I., Ortiz, J., Hermosin-Gutierrez, I., Urretavizcaya,I. et al., Color, ellagitannins, anthocyanins, and antioxidantactivity of Andean blackberry (Rubus glaucus Benth.) wines.J. Agric. Food Chem. 2012, 60, 7463–7473.

[45] Grussu, D., Stewart, D., McDougall, G. J., Berry polyphenolsinhibit �-amylase in vitro: Identifying active components inrowanberry and raspberry. J. Agric. Food Chem. 2011, 59 ,2324–2331.

[46] Boath, A. S., Stewart, D., McDougall, G. J., Berry componentsinhibit �-glucosidase in vitro: synergies between acarboseand polyphenols from black currant and rowanberry. FoodChem. 2012, 135, 929–936.

[47] Kahle, K., Kraus, M., Scheppach, W., Ackermann, M. et al.,Studies on apple and blueberry fruit constituents: do thepolyphenols reach the colon after ingestion? Mol. Nutr. FoodRes. 2006, 50, 418–423.


research article anthocyanins and proanthocyanidins …...anc and pac (100 m c3g and epicatechin...

Documents