Fitoterapia 97 (2014) 78–86

Contents lists available at ScienceDirect

Fitoterapia journal homepage: www.elsevier.com/locate/fitote

The antibiofilm activity of lingonberry flavonoids against oral pathogens is a case connected to residual complexity Kaisu R. Riihinen a,b,⁎, Zhen M. Ou c, Tanja Gödecke d, David C. Lankin d, Guido F. Pauli a,d, Christine D. Wu c a b c d

Institute for Tuberculosis Research, College of Pharmacy, University of Illinois at Chicago, 833 S. Wood Street, Chicago, IL 60612-7231, USA Institute of Public Health and Clinical Nutrition, University of Eastern Finland, P.O. Box 1627, 70211 Kuopio, Finland Department of Pediatric Dentistry, College of Dentistry, University of Illinois at Chicago, Chicago, IL 60612-7231, USA Department of Medicinal Chemistry and Pharmacognosy, College of Pharmacy, University of Illinois at Chicago, 833 S. Wood Street, Chicago, IL 60612-7231, USA

a r t i c l e

i n f o

Article history: Received 9 March 2014 Accepted in revised form 18 May 2014 Available online 29 May 2014 Keywords: Flavonoids Polyphenols Dental caries Vaccinium vitis-idaea Streptococcus mutans Fusobacterium nucleatum Chemical compounds studied in this article: Quercetin-3-rhamnoside (PubChem CID: 5280459) Quercetin-3-glucopyranoside (PubChem CID: 5280804) Quercetin-3-arabinofuranoside (PubChem CID: 5490064) Proanthocyanidin A2 (PubChem CID: 124025) Proanthocyanidin A1 (PubChem CID: 9872976) Proanthocyanidin B2 (PubChem CID: 122738) Cyanidin-3-galactoside (PubChem CID: 44256700) Cyanidin-3-glucopyranoside (PubChem CID: 185664) (+)-Catechin (PubChem CID: 9064) (−)-Epicatechin (PubChem CID: 72276)

a b s t r a c t The antimicrobial activity of lingonberry (Vaccinium vitis-idaea L.) was evaluated against two oral pathogens, Streptococcus mutans and Fusobacterium nucleatum. Long-bed gel permeation chromatography (GPC; Sephadex LH-20) yielded purified flavonoids, with the most efficient minimum inhibitory concentrations (MICs) against planktonic cells in the anthocyanin and procyanidin primary fractions against F. nucleatum (63–125 μg/ml) and in the procyanidin rich fraction against S. mutans (16–31 μg/ml). The purified flavonol glycosides and procyanidins inhibited biofilm formation of S. mutans (MICs 16–31 μg/ml), while the corresponding reference compounds showed no activity. Secondary GPC purification yielded flavonol glycosides devoid of antibiofilm activity in the 50% MeOH fraction, while elution with 70% acetone recovered a brownish material with activity against S. mutans biofilm (MIC 8 μg/ml). Even after HPLC-PDA, NMR, and MALDI-TOF analyses, the structural identity of this material remained unknown, while its color and analytical characteristics appear to be consistent with flavonoid oxidation products. © 2014 Elsevier B.V. All rights reserved.

⁎ Corresponding author at: Institute of Public Health and Clinical Nutrition, University of Eastern Finland, P.O. Box 1627, 70211 Kuopio, Finland. Tel.: +358 40 5911425; fax: +358 17 162 792. E-mail address: Kaisu.Riihinen@uef.fi (K.R. Riihinen).

http://dx.doi.org/10.1016/j.fitote.2014.05.012 0367-326X/© 2014 Elsevier B.V. All rights reserved.

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

1. Introduction The natural growing environment of lingonberry (Vaccinium vitis-idaea L.) is restricted to the Northern hemisphere. Nordic countries are known for the traditional consumption of food products prepared from these light red colored berry fruits. As a dietary supplement, lingonberry extract is available world-wide [1]. Flavonoids such as anthocyanins, flavonol glycosides, and proanthocyanidins are currently considered to be the health-promoting constituents of the berries [2]. Among Ericaceous berry fruits, lingonberries exhibit intermediate levels of anthocyanins (810 mg/100 g dry weight [DW]) and flavonol glycosides (82 mg/100 g DW), but high content of procyanidins (240–380 mg/100 g DW) compared to bilberries (Vaccinium myrtillus L.) [3,4]. Lingonberry extracts have been found to inhibit the promotional stage of chemically induced carcinogenesis [5], the intestinal tumorigenesis in multiple intestinal neoplasia [6], the growth of periodontal pathogens [7], the hemagglutination of Escherichia coli [8], the aggregation of oral bacterial cells [9], and the binding of Neisseria meningitidis pili to human epithelial cells [10]. Anticarcinogenic and antiadhesive activities were concentrated in the lingonberry fractions rich of flavonoids [5,9,10]. Antimicrobial and antiplaque activities of plant flavonoids, obtained from a variety of different plant sources, have been demonstrated in many in vitro studies [11–13]. Moreover, there is clinical evidence which demonstrates that the consumption of flavonoid-rich foods or beverages has a benefit to oral health by exhibiting anti-gingivitis and anti-caries properties [14–17]. The purification and structural characterization of berry flavonoids is a prerequisite for the determination of their in vitro biological activity [2]. Reported purification protocols are based on high speed countercurrent chromatography [18], gel permeation chromatography (GPC) [19–22], preparative HPLC [8,23], as well as combinations of these techniques. [5,24,25] However, the purity of the isolates has seldom been addressed and/or related to the purification method, even for the major flavonoids which are readily identified using contemporary NMR spectroscopy. In general, the loss of bioactivity during the final purification steps and the observation of inverse correlations between purity levels and bioactivity are indications of the potential presence of residual complexity, a phenomenon widely observed in natural products research [26]. For example, the abolishment of the anti-adhesion activity of cranberry against P-fimbriated E. coli to the human uroepithelial cell line T24 has been noted during the final reversed phase HPLC purification step [24]. A case of inverse purity-activity relationships (PARs) has been described for the antimycobacterial potential of ursolic acid and other triterpenes [27,28]. Therefore, residual complexity concerns pertain not only to flavonoids but to all associated bioactive compounds purified from plant materials [26,29,30]. Since 1H NMR is a universal detector, recently established 100% quantitative 1H NMR (qHNMR) methods allow effective determination of major and minor impurities in natural products [31]. The authors recently found noticeable discrepancies between the purities of isolated flavonol glycosides measured by HPLC (82–94 area-% at 250 nm) and the corresponding purities determined by qHNMR (52–70%) [21]. Similar forms of static residual complexity have been demonstrated to affect other natural product

79

extracts and isolates such as ginkgo extracts, ginkgolide mixtures [32], and Actaea triterpenes [33]. The aim of the present study was to purify lingonberry flavonoids and to examine their antimicrobial activity against selected oral pathogens associated with dental caries and periodontal disease. For this purpose, repeated long-bed gel permeation chromatographic (GPC) separation on Sephadex LH-20 material was employed for the purification of flavonol glycosides [21], anthocyanins, procyanidins, and flavan-3-ols from lingonberry, as well as to remove the residual complex materials from the isolated flavonol glycosides. The effect of the enriched flavonoids on growth and biofilm formation of S. mutans and F. nucleatum was investigated. Identifications were performed by 1D and 2D NMR, HPLC–PDA, HPLC-IT–TOF-MS, UHPLC–PDA, and MALDI-TOF. Methods for quantification were based on the absorbance in HPLC– PDA and the total integration of 1H NMR spectra. 2. Materials and methods 2.1. Standards Quercetin-3-O-β-galactopyranoside (hyperoside or hyperin) was purchased from Indofine Chemical Co. Inc. and quercetin-3-O-α-rhamnopyranoside (quercitrin), quercetin3-O-β-D-glucopyranoside, quercetin-3-O-α-arabinofuranoside (avicularin) were from ChromaDex Inc. (Santa Ana, CA). The following standards were purchased from Sigma Chemical Co.: cyanidin-3-galactoside (idaein chloride), vanillic acid, p-hydroxybenzoic acid, chlorogenic acid, p-coumaric acid, caffeic acid, ferulic acid, quercetin, (+)-catechin, (−)-epicatechin, and procyanidin B2, while cyanidin-3-glucoside was obtained from Extrasynthese (Genay, France). The structures of commercial materials were verified (1D/2D NMR experiments) and purities determined by 1D qHNMR. 2.2. Fractionation of flavonoids A stepwise fractionation of lingonberry flavonoids was achieved on Diaion reversed-phase HP-20 resin and on Sephadex LH-20 based GPC as described previously [21]. Briefly, juice concentrate was pre-fractionated with a MeOH-water-gradient over HP-20 resin to enrich the flavonoids in fractions (Frs.) 3a–3d and 4 (Fig. 1). The composition of the Frs. was monitored by TLC [21]. Since Frs. 3a–3d displayed distinct spot profiles during TLC screening, they were not combined, even though they all eluted at 50% MeOH. Fr. 4 was released from the resin with 100% MeOH. The subsequent GPC fractionation was performed on a Sephadex LH-20 gel bed in a precolumn (85–95 ml/17–19 cm, I.D. 25 mm), or/and in a main column system (total 910 ml/ 960 cm, I.D. 11 mm). Two portions of Fr. 3a (0.8 and 1.0 g) and one portion of Fr. 3b (1.2 g) were re-chromatographed over the precolumn, eluting first with MeOH (1 ml/min) and finally recovering residual material by elution with 70% acetone. These fractions were combined according to TLC for further antimicrobial assays, compound identification, and quantification. The compounds eluting between 110 and 210 ml (1.3– 2.5 bed volumes [BVs]) were visualized as red spots with vanillin sulfuric acid spray. These proposed flavan-3-ols and procyanidins were further eluted with MeOH (1 ml/min)

80

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

Lingonberry juice concentrate into HP-20 resin

water

50% MeOH

20% MeOH

Fr. 3a against S. m. 125 µg/ml and F. n. 250 µg/ml Fr. 3b-3d against S. m. and F. n. 250 µg/ml

100% MeOH

Fr. 4 against S. m. and F. n. 250 µg/ml

2 x Frs. 3a and 3b into GPC precolumn with 100% MeOH Non phenolic material S. m. and F. n. 1000 µg/ml

Anthocyanins 21% S. m. 250 µg/ml F. n 63 µg/ml

Flavan-3-ols & PC dimers (B-type) 32% with S. m. 250 µg/ml 70% acetone F. n. 125 µg/ml

into GPC precolumn and main column with 100% MeOH

into GPC precolumn and main column with 100% MeOH

Flavan-3-ols S. m. 250 µg/ml S. m. biofilm 16 µg/ml F. n. 125 µg/ml

two PC (B-type) dimers S. m. 125 µg/ml S. m. biofilm 16 µg/ml F. n. 250 µg/ml

PC (A-type) dimers and polymers S. m. 16, 31 µg/ml S. m. biofilm 16 µg/ml F. n. 63, 125 µg/ml

with Flavonol glycosides 70% acetone S. m. 250 µg/ml S. m. biofilm 16, 31 µg/ml F. n. 125, 250 µg/ml

PC polymers S. m. 31 µg/ml S. m. biofilm 16 µg/ml F. n. 125 µg/ml

into GPC precolumn with 50% methanol Repurified flavonol glycosides S. m. 1000 µg/ml S. m. biofilm 1000 µg/ml

with 70% acetone

Brownish components S. m. 125 µg/ml S. m. biofilm 8 µg/ml

Fig. 1. Fractionation scheme of lingonberry concentrate with the composition and minimum antimicrobial concentrations (MIC) of fractions obtained. Gel permeation chromatography (GPC) on a Sephadex LH-20 gel bed in a precolumn (85–95 ml/17–19 cm, I.D. 25 mm), or/and in a main column system (total 910 ml/960 cm, I.D. 11 mm). Abbreviations used: fraction, Fr.; Streptococcus mutans, S. m.; Fusobacterium nucleatum, F. n.; procyanidins, PCs.

through the main GPC column system [21]. The Frs. were evaporated to dryness directly in test tubes using a Savant SpeedVac concentrator (Thermo Fisher Scientific, Waltman, MA) to obtain weights of collected Frs. per 10 ml in tube. The higher volumes were evaporated with a rotary evaporator (Buechi R114, Switzerland). Dried samples were taken from all steps of fractionation for the identification and quantification of flavonoids and for the antimicrobial assays as presented in Fig. 1. Flavonol glycosides were previously collected from the main GPC system and analyzed independently for purity and flavonoid composition [21]. These flavonol glycoside fractions were combined for further purification in a Sephadex precolumn equilibrated with 50% H2O-MeOH. The combined flavonol glycosides from the elution range 1200 to 1290 ml (1.32–1.42 BVs) were loaded on the precolumn and released by stepwise elution of 50% MeOH (total 470 ml, 5.5 BVs), MeOH (100 ml), and 70% acetone (150 ml). This order of eluents partially collapsed the Sephadex LH-20 bed and some materials could not be recovered as indicated by a slight red color which persisted throughout the column. In the second trial, with newly packed GPC column, the flavonol glycosides in the elution range from 1340 to 1410 ml (1.47–1.55 BVs) were eluted with 50% MeOH (total 600 ml, 7.0 BVs), and the subsequently retained material was further eluted with 70% acetone (150 ml). The procedure led to an opal white gel bed which was re-equilibrated and re-used. Solvents were dried to obtain samples for analyses as described above.

2.3. Identification and quantification of flavonoids TLC, spectrophotometry, HPLC-MS-IT-TOF, and HPLC–PDA analyses were performed with the same protocols as previously described [21]. The identification of all flavonoids and the conjugates of phenolic and hydroxycinnamic acids in HPLC– PDA chromatograms was facilitated by the availability of previously analyzed on-line UV/Vis spectral library data for phenolic compounds in Finnish berries [4,25,34,35]. The fractions, which were assumed to consist of polymeric procyanidins, were analyzed with a Voyager-DE PRO high performance bench top matrix-assisted laser desorption ionization/time-of-flight (MALDI-TOF) mass spectrometer (Applied Biosystems, Inc., CA). Samples were dissolved in MeOH at a 2 mg/ml concentration for MALDI-TOF mass spectral analysis. A quantity (20 mg/ml) of 2,5-dihydroxybenzoic acid (99%, Agros Organics) was dissolved in water for the matrix solution. The sample and matrix solutions were mixed at 1/1 (v/v) ratio, spotted onto defined locations on the stainless steel MALDI target plate, dried under vacuum at room temperature, and analyzed immediately. Positive ion spectra (600 summed acquisitions) were acquired in the linear mode. The accelerating voltage was set at 20,000 V, and the grid voltage was set at 94% of the accelerating voltage. The delay time was 175 nsec. The mass range window was between m/z 500–5000. The laser intensity was set manually at 2100 units. The singly charged molecular ions of Angiotensin II (1045.5423 Da) and ACTH fragment 18–39 (2464.1989 Da) were used as external

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

standards for the mass calibration procedure (ProteoMassTM MALDI-MS, Sigma-Aldrich). Standards were diluted in MeOH 10 pmol/μl and spotted similarly as samples. UHPLC–PDA analysis was on Shimadzu Nexera UHPLC system equipped with LC-30AD pump, DGU-20A5 online degasser, SIL-30AC autosampler, CTO-30A column oven and Semi-Micro Flowcell SPD-M20A PDA detector. The temperature of column oven and PDA detector was set at 40 °C. The separation was achieved on a 1.7 μm particle size Waters Acquity UPLC BEH C18, 2.1 mm × 50 mm column with VanGuard precolumn (2.1 mm × 5 mm, the same packing material). A 25 min linear gradient of organic phase (B, 0.1% formic acid (FA) in MeCN) in an acidified water phase (A, H2O with 0.1% FA) was run at a flow rate of 0.6 ml/min from 5% to 98%. The organic phase concentration was hold at 98% in 3 min, and finally returned to initial condition in 0.5 min for re-equilibration for 2.2 min. NMR data (400 MHz) of commercial reference compounds were obtained on a Bruker AVANCE-400 spectrometer equipped with a 5-mm broadband ATM probe with the probe temperatures maintained at 25 °C (298 K). The commercial reference compounds were dissolved in MeOH-d4 (99.8% D; Cambridge Isotope Laboratories, Inc., MA) and transferred into the NMR tubes (5 mm NMR tubes; Norell, Candisville, NJ). The total filling volume was adjusted to 600 μl for MeOH-d4. For quantitative 400 MHz 1H NMR (qHNMR) analysis, 13C GARP decoupling was employed for removal of 13C satellites in the 1H spectra [36]. Post-acquisition data processing and qHNMR analysis was performed using NUTS (Acorn NMR Inc., CA). Spectra were referenced to the residual solvent signal, MeOD-d2, at 3.310 ppm. The window function employed used Gaussian line resolution enhancement (line-broadening factor, lb = − 0.3 Hz/Gaussian broadening, gb = 0.05). The number of data points after zero-filling (SI) was 256 K to increase the overall digital resolution of the spectra. The purity of the isolated compounds was determined by the modified 100% method [21]. Briefly, the cleanest signal (no spectral overlap present) was selected for integration and the signal to noise ratio was ≥730 in every case. The total compound integral was calculated by multiplying the selected proton signal by its total number of protons (Hs). The resulting total integral was related to the integral of the entire spectrum (range: ~0.5–10.5 ppm) to yield the content of the compound calculated as integral-% (%Int/%Int), which represents the purity of the compounds (Table 1S in Supplementary Data). The total spectral integral excluded the residual solvent signals (HDO and MeOD-d2). 2.4. Test bacteria and growth conditions F. nucleatum ATCC 10953 and S. mutans 25175 were used in this study. F. nucleatum ATCC 10953 was grown anaerobically in the Schaedler broth (OXOID LTD., Basingstoke, Hampshire, England) in an anaerobic chamber (37 °C, 10% H2, 5% CO2, and 85% N2, Forma Scientific, Inc., Marietta, OH, USA). S. mutans 25175 was also grown anaerobically in a chemically defined medium (CDM) or the Brain Heart Infusion broth (BHI; Difco Laboratories, Detroit, MI, USA). 2.5. Inhibition of growth and biofilm formation The MIC of test fractions or samples on growth and biofilm formation were determined using 96-well microtiter plates as

81

previously described [37]. Each well contained washed cells of F. nucleatum (5 × 106 CFU/ml) or S. mutans (5 × 105 CFU/ml), two-fold diluted test agent and respective growth medium. Controls included inoculated growth medium without test agents, chlorhexidine gluconate (CHX), MeOH and DMSO. All plates were incubated anaerobically for 48 h and growth was determined spectrophotometrically using a microplate reader. MIC is defined as the minimum concentration at which no growth was observed compared to the control. Only S. mutans was used for biofilm inhibition assay, 1% sucrose was added to the growth medium using previously established method [37]. 3. Results and discussion 3.1. Composition of the lingonberry Fr. 3 and Fr. 4 Lingonberry juice concentrate was prefractionated over reversed-phased resin into fractions enriched with the three subclasses of flavonoids (Fig. 1) [21]. Anthocyanins were predominant in Frs. 3a and 3b (10% and 7%, w/w, respectively), while hydroxycinnamic acid and phenolic acid conjugates dominated Frs. 3c and 3d (Fig. 2). Fr. 4 was rich in flavonol glycosides (10%, w/w) and hydroxycinnamic acid (HCA) conjugates [21]. Similarly as HCA conjugates, procyanidins and flavan-3-ols were distributed in variable amounts into Frs. 3a–d and 4. 3.2. Composition of subsequent GPC fractions of Frs. 3a–d Two portions of Fr. 3a and one portion of Fr. 3b were selected for further GPC precolumn fractionation to anthocyanins, flavan-3-ols, and procyanidins (Fig. 1). Fig. 3 shows both the graphic profiles of tube wise weights in the course of GPC elution and the composition of combined fractions as HPLC-PDA profiles at 280 nm. The GPC elution range from 70 to 110 ml (0.8– 1.3 BVs) consisted of anthocyanins 21% (w/w). Major anthocyanins were cyanidin-hexoside and -pentoside identified as positive molecular ions at m/z 449.35 and 419.35 by HPLC-MS, respectively. This is consistent with cyanidin-3-galactoside (88%) and cyanidin-3-arabinoside (11%) being reported as the predominant anthocyanins present in lingonberry [38]. The continuous MeOH flow (110 ml–210 ml, 1.3–2.4 BVs) released (+)-catechin and (−)-epicatechin (protonated ions at m/z 291.25 in HPLC-MS), B-type dimers (at m/z 579.45–579.50) (Fig. 3), and B-type trimers (at m/z 867.50–867.60) from the GPC precolumn. Structural characteristics of A- and B-type procyanidins are presented in Fig. 4. There was also a hump like drift in a HPLC-PDA chromatographic baseline at 280 nm (Fig. 3). As a next purification step for the present fraction, a 10 meter GPC purification separated flavan-3-ols (1.68 to 1.77 BV) and B-type procyanidin dimers (1.88 to 2.07 BV) into separate fractions and off from other materials (Fig. 1). Subsequent elution with 70% acetone recovered residual material from GPC precolumn. Analytically, this material consisted of two distinctive peaks of A-type procyanidin dimers (at m/z 577.40–577.45) (Fig. 3) and a distinctive hump in the HPLC-PDA chromatographic profile. This fraction was further analyzed with MALDI-TOF, as high molecular weight procyanidins were expected to be present in the hump with the same absorption spectra as flavan-3-ols [22]. Procyanidins ionized as sodium adducts (+22.99 Da) at m/z

82

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

AC

mAU AC

40 20

mAU

Fr. 3a

Un

PA

10

40

AC

20

AC/PA 15 AC/HCA

Un

PA

mAU

15

PA

20

40

PA

20 5

25

15 AC/HCA HCA HCA PC

10

min

Fr. 3c Un PC

PC 10

mAU

20

HCA AC/HCA

Un

PA

Fr. 3b Un PC

PC

10

60 40

min

25

20

Un

PC

20 HCA Un

15

25 PC FG

min

Fr. 3d

25

20

min

Fig. 2. HPLC-PDA chromatograms of lingonberry Frs. 3a–d (Fig. 1.) recorded at 280 nm for equivalent concentration of samples (1 mg/ml). Peaks with the typical on-line UV/Vis spectrum are identified as previously [34,35]. Abbreviations used: phenolic acid conjugate, PA; anthocyanin, AC; hydroxycinnamic acid conjugate, HCA; procyanidin, PC; (+)-catechin, (+)-C; (−)-epicatechin, (−)-EC; unidentified, Un; flavonol glycoside, FG.

600.08, 888.10, 889.96, 1176.38, 1178.28, 1464.35, 1466.91 and 1753.85 (Fig. 1S in Supplementary Data). These ions correspond to the calculated formula weights of procyanidin structures: dimers A (576.50 Da), trimers A and B (864.76 and 866.78 Da), tetramers A and B (1153.02 Da and 1155.04 Da), pentamer A (1441.28 Da), and hexamer A (1729.54 Da) (Table 2S in Supplementary Data), respectively.

was previously described [21,39]. After the flavonol glycosides had entered the main column system, the precolumn was disconnected and the highly retained material eluted from the precolumn with 70% acetone. MALDI-TOF analysis afforded the following major ions occurring as sodium adducts at m/z 600.44, 888.80, 890.71, 1177.52, and 1464.83, consistent with the formula weights of dimer A, trimers A and B, tetramer A, and pentamer A, respectively.

3.3. Composition of Fr. 4 3.4. Antimicrobial activity against planktonic cells The distribution of flavonol glycosides upon fractionation from 1.3 to 1.8 BVs on the main GPC column system and the subsequent identification from the fingerprint NMR spectra

All Frs. 3a–d and Fr. 4 showed moderate MICs (125 and 250 μg/ml) against the growth of F. nucleatum and S. mutans,

mg 0

100 200 300

10

Cy-hex

methanol 1 ml/min mAU

60

Cy-pent

30

AC AC

PA

mAU

10

HCA

15

PA

20

25

20 PA

10

DiB Un

methanol 110-210 ml

(-)-EC DiB HCA TriB DiB

Un

70% acetone

20

10

mAU

15

DiA

20

ml

DiA

25

min

70% acetone 0-100 ml

10 70

min

(+)-C

30 160

methanol 70-110 ml

Cy-hex

10 110

cyanidin

50

5

hump

5

10

15

20

25

min

Fig. 3. Graphic presentation of GPC separations (Sephadex LH-20, 85 ml) performed for two portions of Fr. 3a (■, ) and one portion of Fr. 3b (•). Composition of combined GPC fractions is presented as HPLC-PDA chromatograms recorded at 280 nm for equivalent concentrations of samples (1 mg/ml). See Fig. 2 for abbreviations used in the peak labels.

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

83

(MICs 125–250 μg/ml). Two lingonberry flavonol glycosides from Fr. 4, quercetin-3-O-[4″-(3-hydroxy-3-methylglutaroyl)]α-rhamnopyranoside and Q-3-O-α-rhamnopyranoside, were re-purified by running them through the precolumn with 50% MeOH and recovering the retained material with 70% acetone (Fig. 1, Table 2). The recovered material had a brown color and apparently was part of the residual complexity detected by the qHNMR analysis of lingonberry flavonol glycosides [17]. Of these residually complex primary fractions, those with procyanidins were more active (MIC 31 μg/ml, Fig. 1) than those without (MIC 125 μg/ml, Fig. 1, Table 2) when evaluating their growth inhibitory effects towards S. mutans. Fig. 4. Chemical structures of B- and A-type procyanidin dimers.

which warranted further fractionation and purification steps (Fig. 1). This amount of flavonoids may be obtained from one teaspoonful of fresh lingonberries (3 g) consisting approximately of 400 μg of flavonol glycosides, 4000 μg of anthocyanins, and 750 μg of flavan-3-ols [4]. Frs. enriched with anthocyanins, flavan-3-ols and procyanidins inhibited the growth of F. nucleatum at 63–125 μg/ml, consistent with concentration effects of the activity. However, the antimicrobial activity against F. nucleatum could not be further concentrated in the following fractionation steps. Lingonberry flavan-3-ols and procyanidins dimers inhibited growth of planktonic S. mutans cells at MIC ranging 125–250 μg/ml. However, the final elution of the GPC column with 70% acetone recovered a brownish colored material exhibiting increased growth inhibition against S. mutans (MIC 16, 31 μg/ml). Interestingly, commercially available reference materials of flavonol glycosides (yellow, qHNMR purity 86.1–97.2%) and (+)-catechin exhibited no activity in the assay (MIC N 500 μg/ml; Table 1), while the purified lingonberry flavonol glycosides (brownish yellow, 52–70% purity by qHNMR analysis) exhibited modest antimicrobial activity against the growth of planktonic F. nucleatum and S. mutans

3.5. Antibiofilm activity of lingonberry constituents GPC purified lingonberry flavan-3-ols and procyanidins dimers were more active against the resistant biofilm cells (MIC 16 μg/ml) than against planktonic cells (MIC 125–250 μg/ml) (Table 1, Fig. 1). However, the commercially available (+)-catechin (97.2% purity by qHNMR analysis) was shown to be inactive (Table 1). Similarly, lingonberry flavonol glycosides (52–70% purity by qHNMR analysis) inhibited S. mutans biofilm formation (MIC 16–31 μg/ml), whereas the reference compounds (86.1–97.2% qHNMR purity) did not exhibit any activity (Table 2). A brown material with antibiofilm activity was enriched from Fr. 3 (Fig. 1), Fr. 4 (Fig. 1), and lingonberry flavonol glycosides (Table 2) by GPC eluting with 70% acetone. All the antibiofilm activity (MIC 8 μg/ml) was in the brownish residually complex material, while re-purified flavonol glycosides were essentially inactive, as were the commercial standards. The present fractionation protocol did not separate A-type dimers and polymeric procyanidins from the brownish residually complex material, as both apparently co-eluted from the Sephadex LH-20 material, despite the very long GPC system used. Therefore, the role of procyanidins in antibiofilm activity may even be negligible, since equivalent activity was found for the brownish fractions with and without procyanidins (Fig. 1).

Table 1 Antimicrobial activity (MIC μg/ml) and purity of purified lingonberry flavonoids and commercial quercetin glycosides and (+)-catechin. The structures of Q-3-O-[4″-(3-hydroxy-3-methylglutaroyl)]-α-rhamnopyranoside, Q-3-O-α-rhamnopyranoside, Q-3-O-β-galactopyranoside, Q-3-O-β-glucopyranoside, and Q-3-O-α-arabinofuranoside were identified by NMR as reported previously [39]. Fractions (HPLC puritya)

MeOH bed volumes

qHNMRb

F. n.c

S. m.

S. m. biofilms

GPC purified (100% MeOH) Q-HMG-rhap (81 area-%) Q-HMG-rhap (87 area-%) Q-rhap (85 area-%) Q-rhap (82 area-%) Q-galp (75 area-%) Q-galp (94 area-%) Q-galp (90 area-%) Flavan-3-ols (50 area-%) PC B-type dimers (32 area-%)

1.35–1.36 1.36–1.37 1.51–1.52 1.53–1.54 1.57–1.58 1.58–1.59 1.59–1.62 1.68–1.77 1.88–2.07

NA 67 NA 70 NA 52 NA NA NA

250

250

31

250

250

31

125

250

16

125 125 250

250 250 125

16 16 16

96.4 96.5 95.6 86.1 97.2

N500 NA N500 NA NA

N500 N500 N500 N500 1000

N1000 N1000 N1000 N1000 1000

Commercial flavonoids Q-galp Q-araf Q-rhap Q-glcp (+)-catechin a b c

HPLC-PDA purity as area-% with UV-detection at 250 nm (Q-gly) and 254 nm (Flavan-3-ols and procyanidins). qHNMR purity was assessed previously for adjacent fractions with total integral and the modified 100% method [21]. Abbreviations used: Streptococcus mutans, S. m.; Fucobacterium nucleatum, F. n.; not analyzed, NA, quercetin, Q; procyanidin, PC.

84

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

Table 2 Antimicrobial activity (MIC) of re-purified quercetin-3-O-[4″-(3-hydroxy3-methylglutaroyl)]-α-rhamnopyranoside (Q-HMG-rhap) and Q-3-O-αrhamnopyranoside (Q-rhap) eluted out of precolumn GPC running 50% H2O-MeOH. Fractions Original Q-HMG-rhap 60–250 ml 250–370 ml 370–410 ml 410–470 ml MeOH 70% acetone in H2O Original Q-rhap 60–340 ml 340–410 ml 410–500 ml 500–680 ml 70% acetone in H2O (Fig. 5)

Weight/mg

% of total weight

S. m.

S. m. biofilms

100

250

31

8.5 13.2 2.7 10.0 4.4 2.4

21 32 7 24 11 6

500 1000 1000 1000 250 125

500 1000 1000 1000 16 8

8.9 4.7 12.4 2.1 2.6

100 29 15 40 7 8

125 125 250 500 250 125

31 125 125 250 125 b16

UV 80000

Quercetin

60000

40000

Q-rhap

20000

Moreover, it remained unresolved whether the brownish material was responsible for antibiofilm activity or it activated flavonol glycosides and flavan-3-ols. Compounds with antibiofilm activity are either an anti-adhesive physical barrier and/or inactivate essential enzymes of oral biofilm formation. Previously, we have found that the tea flavonoid (epigallocatechin gallate [95% purity by HPLC] EGCg) inhibits S. mutans biofilm formation at MIC 16 μg/ml [12,13]. Detailed analysis showed that EGCg significantly inhibits the gtf genes associated with biofilm formation [13]. The activity was associated with the pyrogallol structural motifs. Other postulated mechanisms against bacterial biofilms have been presented for cranberry procyanidins [40]. Berry procyanidins may prevent aggregation, reduce bacterial hydrophobicity or alter cell surface molecules [40]. 3.6. The identity of the brownish residually complex material In Sephadex LH-20 GPC, the active brownish material co-eluted with flavonol glycosides, flavan-3-ols, and procyanidins when using neat MeOH, and also during subsequent elution with 70% acetone. Apparently the residually complex material was loosely “bound” to flavonoids in this range of GPC polarity and molecule size. However, the brownish material could be mostly separated from the flavonol glycosides by replacing MeOH as an eluent with the more polar 50% MeOH in water. While this led to a further purification of the flavonol glycosides, the brownish material with minor amounts of quercetin and Q-3-O-α-rhamnopyranoside was subsequently eluted with 70% acetone (Fig. 5). The identity of the brownish material was studied with HPLC-PDA, NMR, and MALDI-TOF. It gave an unresolved hump in UHPLC-PDA, with on-line light absorption spectra similar to procyanidins [25], and also gave rise to broad resonance peaks in the 1H NMR spectrum (5.8–7.7 ppm) (Fig. 5). Such broad signals in the aromatic region of NMR spectra are typical ascribed to high molecular weight procyanidins [22]. The unresolved humps and a drift of

2.0

4.0

6.0

8.0

10.0

5’

12.0

14.0 min

Broad resonances

Quercetin Q-rhap

8 6

2’ 6’ 2’&6’

9.0

8.5

8.0

7.5

7.0

6.5

6.0

5.5

ppm

Fig. 5. The brownish component isolated from residually complex Q-3-O-α-rhamnopyranoside (Table 2) above as an UHPLC-PDA chromatogram at 254 nm compared to methanol baseline and below as a proton resonance NMR spectrum 1H at 400 MHz.

the baseline in the RP chromatogram have been broadly investigated and previously described for polymeric procyanidins [41]. However, similar chromatographic behavior was also demonstrated for oxidized procyanidins found in cider [42] and the arubigins found in fermented teas [43]. The Gaussian shaped hump in RP-HPLC indicates the presence of compounds with an intermediately hydrophobic structure [43]. No procyanidin ions were detected by MALDI-TOF MS (Fig. 3S in Supplementary Data). Instead, the spectrum showed unfamiliar ions appearing at m/z 551.83, 574.52, 662.87 and 750.96. This suggests that the compounds responsible for the residual complexity of the (impure) bioactive flavonoids were not procyanidins, but rather may result from oxidation of or association with the flavonoids. The brown color is consistent with quinone-type oxidation products of flavonoids and is a well-known characteristic of fermented tea, in which a brownish color is formed in the course of oxidation. The formation of a brown color in plants is also associated with flavonoids that undergo oxidation to quinones structures, which may associate (complex) reversibly or irreversibly with amino acids and proteins in crushed plant material during harvest or disintegrating cells [44]. Interestingly, earlier work has discussed the formation of quinone-like oxidative products of flavonoids as being partly responsible for the protein-binding and enzyme inactivating properties of flavonoids [45]. Similarly, quinones were identified as active intermediates of flavonoids in some of

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86

their biological activities [46,47]. Overall, the detailed identification concerning the nature of the chemistry that underlies the residual complexity of plant flavonoids still remains unclear, but very likely is related to oxidative processes.

3.7. The structure-activity relationship of lingonberry flavonoids Previously, antiaggregation activity against oral bacteria has been concentrated in flavonoid rich fractions of lingonberries [9]. Moreover, purified A-type procyanidin trimers that were identified by NMR exhibited a strong antimicrobial activity against Porphyromonas gingivalis and Prevotella intermedia [7]. In the present study, the antimicrobial activity against planktonic cells was associated to both the polymeric and A-type lingonberry procyanidins, whereas antibiofilm activity was found to arise from the unidentified brownish and residually complex material. However, interpretation of results was hampered for the procyanidin fractions as the brownish material could not be separated even when using a 10 meter GPC column. Similarly, GPC on Sephadex with MeOH elution did not separate flavonol glycosides from brown colored impurities, but separation was possible when eluting with 50% H2O-MeOH. Presumably MeOH-water releases weak interactions between the flavonol glycoside and impurities. Due to similar elution characteristics, the separation of brown impurities from the A-type dimeric and trimeric to polymeric procyanidins is a more challenging purification task. Synthesis of procyanidins as a means of establishing structure-activity relationships was beyond the scope of this study. However, based on the present observations, interpretations ascribing bioactivities to A-type and polymeric procyanidins should be approached with caution as long as a brown residually complex color appears in the investigated fractions and purified compounds. In fact, the appearance of a residual brown color is a rapid way to distinguish impure from pure procyanidin materials. It is anticipated that the 100% qHNMR method can be tailored further to quantify the polymeric procyanidins within brown colored, impure fractions. The apparently ubiquitous impurities of procyanidins may be overcome by developing more effective separation techniques, combined with a simultaneous monitoring of the purity of fractions by qHNMR.

Abbreviations CFU colon forming unit DW dry weight GPC long-bed gel permeation chromatography MIC minimum inhibitory concentration

Acknowledgments The authors wish to thank Dr. M. Florencia Rodriguez Brasco for her assistance with the UHPLC analyses. This study was financially supported by the Academy of Finland (Post-doctoral researcher's project 127340).

85

Appendix A. Supplementary data The summary of the qHNMR analyses using the modified 100% method; the original MALDI-TOF spectra; calculated molecular formula weights of procyanidins. This paper represents part 23 of the series on Residual Complexity and Bioactivity. This information is available free of charge via the Internet at http://pubs.acs.org. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.fitote.2014.05.012. References [1] Mane C, Loonis M, Juhel C, Dufour C, Malien-Aubert C. Food grade lingonberry extract: polyphenolic composition and in vivo protective effect against oxidative stress. J Agric Food Chem 2011;59:3330–9. [2] Espín JC, García-Conesa MT, Tomás-Barberán FA. Nutraceuticals: facts and fiction. Phytochemistry 2007;68:2986–3008. [3] Hellström JK, Törrönen AR, Mattila PH. Proanthocyanidins in common food products of plant origin. J Agric Food Chem 2009;57:7899–906. [4] Määttä-Riihinen KR, Kamal-Eldin A, Mattila PH, Gonzalez-Paramas AM, Törrönen AR. Distribution and contents of phenolic compounds in eighteen Scandinavian berry species. J Agric Food Chem 2004;52:4477–86. [5] Bomser J, Madhavi DL, Singletary K, Smith MA. In vitro anticancer activity of fruit extracts from Vaccinium species. Planta Med 1996;62:212–6. [6] Misikangas M, Pajari AM, Päivärinta E, Oikarinen SI, Rajakangas J, Marttinen M, et al. Three Nordic berries inhibit intestinal tumorigenesis in multiple intestinal neoplasia/+ mice by modulating beta-catenin signaling in the tumor and transcription in the mucosa. J Nutr 2007;137:2285–90. [7] Ho KY, Tsai CC, Huang JS, Chen CP, Lin TC, Lin CC. Antimicrobial activity of tannin components from Vaccinium vitis-idaea L. J Pharm Pharmacol 2001;53:187–91. [8] Kylli P, Nohynek L, Puupponen-Pimiä R, Westerlund-Wikstrom B, Leppänen T, Welling J, et al. Lingonberry (Vaccinium vitis-idaea) and European cranberry (Vaccinium microcarpon) proanthocyanidins: isolation, identification, and bioactivities. J Agric Food Chem 2011;59:3373–84. [9] Riihinen K, Ryynänen A, Toivanen M, Könönen E, Törrönen R, TikkanenKaukanen C. Antiaggregation potential of berry fractions against pairs of Streptococcus mutans with Fusobacterium nucleatum or Actinomyces naeslundii. Phytother Res 2011;25:81–7. [10] Toivanen M, Ryynänen A, Huttunen S, Duricova J, Riihinen K, Törrönen R, et al. Binding of Neisseria meningitidis pili to berry polyphenolic fractions. J Agric Food Chem 2009;57:3120–7. [11] Furiga A, Lonvaud-Funel A, Dorignac G, Badet C. In vitro anti-bacterial and anti-adherence effects of natural polyphenolic compounds on oral bacteria. J Appl Microbiol 2008;105:1470–6. [12] Xu X, Zhou XD, Wu CD. The tea catechin epigallocatechin gallate suppresses cariogenic virulence factors of Streptococcus mutans. Antimicrob Agents Chemother 2011;55:1229–36. [13] Xu X, Zhou XD, Wu CD. Tea catechin epigallocatechin gallate inhibits Streptococcus mutans biofilm formation by suppressing gtf genes. Arch Oral Biol 2012;57:678–83. [14] Ferrazzano GF, Amato I, Ingenito A, De Natale A, Pollio A. Anticariogenic effects of polyphenols from plant stimulant beverages (cocoa, coffee, tea). Fitoterapia 2009;80:255–62. [15] Wu CD. The impact of food components and dietary factors on oral health. J Food Drug Anal 2012;20(Suppl. 1):270–4. [16] Wu CD, Wei GX. Tea as a functional food for oral health. Nutrition 2002;18:443–4. [17] Signoretto C, Canepari P, Stauder M, Vezzulli L, Pruzzo C. Functional foods and strategies contrasting bacterial adhesion. Curr Opin Biotechnol 2012;23:160–7. [18] Valls J, Millan S, Marti MP, Borras E, Arola L. Advanced separation methods of food anthocyanins, isoflavones and flavanols. J Chromatogr A 2009;1216:7143–72. [19] Morimoto S, Nonaka GI, Nishioka I. Tannins and related-compounds. LX. Isolation and characterization of proanthocyanidins with a doubly-linked unit from Vaccinium vitis-idaea L. Chem Pharm Bull 1988;36:33–8. [20] Kandil FE, Smith MA, Rogers RB, Pepin MF, Song LL, Pezzuto JM, et al. Composition of a chemopreventive proanthocyanidin-rich fraction from cranberry fruits responsible for the inhibition of 12-O-tetradecanoyl

86

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

K.R. Riihinen et al. / Fitoterapia 97 (2014) 78–86 phorbol-13-acetate (TPA)-induced ornithine decarboxylase (ODC) activity. J Agric Food Chem 2002;50:1063–9. Riihinen KR, Gödecke T, Pauli GF. Purification of berry flavonol glycosides by long-bed gel permeation chromatography. J Chromatogr A 2012;1244:20–7. Schmidt BM, Howell AB, McEniry B, Knight CT, Seigler D, Erdman Jr JW, et al. Effective separation of potent antiproliferation and antiadhesion components from wild blueberry (Vaccinium angustifolium Ait.) fruits. J Agric Food Chem 2004;52:6433–42. Ek S, Kartimo H, Mattila S, Tolonen A. Characterization of phenolic compounds from lingonberry (Vaccinium vitis-idaea). J Agric Food Chem 2006;54:9834–42. Turner A, Chen SN, Nikolic D, van Breemen R, Farnsworth NR, Pauli GF. Coumaroyl iridoids and a depside from cranberry (Vaccinium macrocarpon). J Nat Prod 2007;70:253–8. Määttä-Riihinen KR, Kähkönen MP, Törrönen AR, Heinonen IM. Catechins and procyanidins in berries of Vaccinium species and their antioxidant activity. J Agric Food Chem 2005;53:8485–91. Pauli GF, Chen SN, Friesen JB, McAlpine JB, Jaki BU. Analysis and purification of bioactive natural products: the AnaPurNa study. J Nat Prod 2012;75:1243–55. Jaki BU, Franzblau SG, Chadwick LR, Lankin DC, Zhang F, Wang Y, et al. Purity-activity relationships of natural products: the case of anti-TB active ursolic acid. J Nat Prod 2008;71:1742–8. Qiu F, Cai G, Jaki BU, Lankin DC, Franzblau SG, Pauli GF. Quantitative purity-activity relationships of natural products: the case of antituberculosis active triterpenes from Oplopanax horridus. J Nat Prod 2013;76:413–9. Chen SN, Lankin DC, Chadwick LR, Jaki BU, Pauli GF. Dynamic residual complexity of natural products by qHNMR: solution stability of desmethylxanthohumol. Planta Med 2009;75:757–62. Qiu F, Imai A, McAlpine JB, Lankin DC, Burton I, Karakach T, et al. Dereplication, residual complexity, and rational naming: the case of the Actaea triterpenes. J Nat Prod 2012;75:432–43. Gödecke T, Yao P, Napolitano JG, Nikolic D, Dietz BM, Bolton JL, et al. Integrated standardization concept for Angelica botanicals using quantitative NMR. Fitoterapia 2012;83:18–32. Napolitano JG, Gödecke T, Rodriguez-Brasco MF, Jaki BU, Chen SN, Lankin DC, et al. The tandem of full spin analysis and qHNMR for the quality control of botanicals exemplified with Ginkgo biloba. J Nat Prod 2012;75:238–48.

[33] Qiu F, Friesen JB, McAlpine JB, Pauli GF. Design of countercurrent separation of Ginkgo biloba terpene lactones by nuclear magnetic resonance. J Chromatogr A 2012;1242:26–34. [34] Määttä KR, Kamal-Eldin A, Törrönen AR. High-performance liquid chromatography (HPLC) analysis of phenolic compounds in berries with diode array and electrospray ionization mass spectrometric (MS) detection: Ribes species. J Agric Food Chem 2003;51:6736–44. [35] Määttä-Riihinen KR, Kamal-Eldin A, Törrönen AR. Identification and quantification of phenolic compounds in berries of Fragaria and Rubus species (family Rosaceae). J Agric Food Chem 2004;52:6178–87. [36] Pauli GF, Jaki BU, Lankin DC. A routine experimental protocol for qHNMR illustrated with Taxol. J Nat Prod 2007;70:589–95. [37] Xu X, Zhou XD, Wu CD. Tea catechin EGCg suppresses the mgl gene associated with halitosis. J Dent Res 2010;89:1304–8. [38] Andersen ØM. Chromatographic-separation of anthocyanins in cowberry (lingonberry) Vaccinium-vites-idaea L. J Food Sci 1985;50:1230–2. [39] Riihinen KR, Mihaleva VV, Gödecke T, Soininen P, Laatikainen R, Vervoort JM, et al. 1H NMR fingerprinting of Vaccinium vitis-idaea flavonol glycosides. Phytochem Anal 2013;24:476–83. [40] Feghali K, Feldman M, La VD, Santos J, Grenier D. Cranberry proanthocyanidins: natural weapons against periodontal diseases. J Agric Food Chem 2011;60:5728–35. [41] Rohr GE, Meier B, Sticher O. Analysis of procyanidins. New York: Elsevier Science B.V.; 2000. [42] Lea AGH. Reversed-phase gradient high-performance liquid chromatography of procyanidins and their oxidation products in ciders and wines, optimised by Snyder's procedures. J Chromatogr A 1980;194:62–8. [43] Kühnert N, Clifford MN, Muller A. Oxidative cascade reactions yielding polyhydroxy-theaflavins and theacitrins in the formation of black tea thearubigins: evidence by tandem LC-MS. Food Funct 2010;1:180–99. [44] Bittner S. When quinones meet amino acids: chemical, physical and biological consequences. Amino Acids 2006;30:205–24. [45] Kühnau J. The flavonoids. In: Bourne GH, editor. A class of semiessential food components: their role in human nutrition. World Rev Nutr DietBasel, Switzerland: S. Krager; 1976. p. 117–91. [46] Muzolf-Panek M, Gliszczynska-Swiglo A, de Haan L, Aarts JM, Szymusiak H, Vervoort JM, et al. Role of catechin quinones in the induction of EpRE-mediated gene expression. Chem Res Toxicol 2008;21:2352–60. [47] Tu T, Giblin D, Gross ML. Structural determinant of chemical reactivity and potential health effects of quinones from natural products. Chem Res Toxicol 2011;24:1527–39.